Mitochondrial cytochromes c and c1 are core components of the respiratory chain of all oxygen-respiring eukaryotes. These proteins contain haem, covalently bound to the polypeptide in a catalysed post-translational modification. In all eukaryotes, except members of the protist phylum Euglenozoa, haem attachment is to the cysteine residues of a CxxCH haem-binding motif. In the Euglenozoa, which include medically relevant trypanosomatid parasites, haem attachment is to a single cysteine residue in an AxxCH haem-binding motif. Moreover, genes encoding known c-type cytochrome biogenesis machineries are all absent from trypanosomatid genomes, indicating the presence of a novel biosynthetic apparatus. In the present study, we investigate expression and maturation of cytochrome c with a typical CxxCH haem-binding motif in the trypanosomatids Crithidia fasciculata and Trypanosoma brucei. Haem became attached to both cysteine residues of the haem-binding motif, indicating that, in contrast with previous hypotheses, nothing prevents formation of a CxxCH cytochrome c in euglenozoan mitochondria. The cytochrome variant was also able to replace the function of wild-type cytochrome c in T. brucei. However, the haem attachment to protein was not via the stereospecifically conserved linkage universally observed in natural c-type cytochromes, suggesting that the trypanosome cytochrome c biogenesis machinery recognized and processed only the wild-type single-cysteine haem-binding motif. Moreover, the presence of the CxxCH cytochrome c resulted in a fitness cost in respiration. The level of cytochrome c biogenesis in trypanosomatids was also found to be limited, with the cells operating at close to maximum capacity.
Cytochrome c is a key component of the mitochondrial respiratory chain of all oxygen-respiring eukaryotes. Its principal role is electron transfer from cytochrome c1 of ubiquinol:cytochrome c oxidoreductase (Complex III) to the CuA centre of cytochrome aa3 oxidase (Complex IV). c-Type cytochromes are widespread in bacteria, archaea, mitochondria and chloroplasts [1–4]. They are characterized by covalent attachment of haem (Fe-protoporphyrin IX) to protein. This haem attachment is by thioether bonds almost always formed between the vinyl groups of the haem and the two cysteine residues of a CxxCH haem-binding motif (Figure 1); the histidine residue becomes an axial ligand to the haem iron atom and the ‘x’ residues can be virtually any amino acids, except cysteine.
Haem attachment to c-type cytochromes
The covalent haem attachment in cytochromes c has strictly conserved regio- and stereo-specificity [1,5], and, perhaps surprisingly, variants of three distinct complexes are known to catalyse this post-translational modification in different organisms and organelles (reviewed in [6–10]). Uniquely, however, protists from the phylum Euglenozoa, including trypanosome and Leishmania parasites, contain mitochondrial cytochromes c and c1 with haem attached to a single cysteine residue [an A(A/G)QCH haem-binding motif in cytochrome c and FAPCH in cytochrome c1] (Figure 1) [11–15]. Moreover, within the completed nuclear and mitochondrial genome sequences of multiple trypanosomatid species there is no trace of any component from the known c-type cytochrome biogenesis machineries . Thus, in addition to exceptional ‘single cysteine’ cytochromes c, trypanosomatid mitochondria are likely to have a novel cytochrome c biogenesis apparatus. Phylogenetic and biochemical analyses indicate that this machinery is likely to be spread throughout, but confined to, the Euglenozoa .
The Euglenozoa comprises three groups of protists: the little-studied diplonemids which are mostly marine flagellates; the euglenids, a diverse assortment of free-living phagotrophs, osmotrophs and photoautotrophs that include Euglena gracilis; and the kinetoplastids, including the parasitic trypanosomatid family [16,17]. The trypanosomatids incorporate the African sleeping sickness parasite Trypanosoma brucei, the American trypanosome T. cruzi (causal agent of Chagas’ disease), pathogenic Leishmania species and Crithidia species (parasites of arthropods). Many aspects of mitochondrial function in trypanosomatids and other euglenozoans are unique, notably mitochondrial genome architecture and uracil insertion/deletion RNA editing [18–20]. However, the occurrence of single-cysteine mitochondrial cytochrome c only in euglenozoans, and the likelihood of a novel cytochrome c biogenesis apparatus, are peculiarities pertaining to a fundamental and otherwise conserved aspect of mitochondrial function, the assembly and function of the respiratory chain. It is not clear why single-cysteine cytochrome c evolved in the ancestor of extant euglenozoans. The structure of cytochrome c from the trypanosomatid Crithidia fasciculata is remarkably similar to that of a typical mitochondrial cytochrome c (e.g. from yeast and animals), other than the missing thioether bond . A CxxCH haem-binding motif variant of T. brucei cytochrome c was matured stably and apparently with two thioether bonds in Escherichia coli by the cytochrome c maturation proteins of that organism . From these results, the euglenozoan cytochrome c proteins clearly could, in principle, accommodate a second thioether bond. We have previously hypothesized (e.g. [11,22]) that the presence of single-cysteine cytochrome c in euglenozoans might reflect a distinctive redox environment in their mitochondrial IMS (intermembrane space) (where cytochrome c is located). For example, it may be that within the IMS of euglenozoan cells a strong oxidant would irreversibly oxidize the cysteine residues of a typical CxxCH apocytochrome c to a disulfide bond. Single-cysteine cytochrome c may provide a means of bypassing such an interaction . Alternatively, the absence of a CxxCH cytochrome c may alleviate constraints on other redox proteins in the IMS. In the present study we have investigated the expression and maturation of cytochrome c with a typical CxxCH haem-binding motif in trypanosome mitochondria.
C. fasciculata choanomastigotes were cultured at 27°C in Warren medium supplemented with heat-inactivated FBS (fetal bovine serum) and haemin, as described previously . Procyclic T. brucei was cultured in SDM-79 medium  supplemented with 10% (v/v) heat-inactivated FBS and haemin (7.5 μg/ml). For incubations with the mitochondrial respiratory chain inhibitors sodium azide and antimycin A, cultures were inoculated at a starting density of 2×106 cells/ml, grown for 24 h in the presence of inhibitor and then counted in duplicate using a Neubauer haemocytometer. Sodium azide (Sigma–Aldrich) was added to cultures at a final concentration of 1 mM (from a stock solution of 0.1 M in water) with lower concentrations achieved by 10-fold serial dilutions. Antimycin A (Sigma–Aldrich) was added to a final concentration of 25 μg/ml (from a stock solution of 25 mg/ml in ethanol) and again lower concentrations were achieved by 10-fold serial dilutions.
Trypanosomatid molecular genetics
A variant of T. brucei cytochrome c with a CAQCH haem-binding motif was cloned from plasmid pKK223-TbcytcCXXCH  into the KpnI and XhoI sites of the C. fasciculata expression vector pNUS-GFPc  creating plasmid pKSH035. The WT (wild-type) haem-binding motif (AAQCH) was restored by QuikChange® site-directed mutagenesis of pKSH035 to produce plasmid pKSH034. For gene disruption of cytochrome c in T. brucei, a blasticidin deaminase gene flanked by tubulin and actin processing signals was PCR-amplified from the pCP101 template  using the primer combination 5′-ATGAACATATCTGAAGAACCGTATTTGTTGGGTCCCATTGTTTGCCTC-3′ and 5′-CTTTAATGTCTCGAGGTATGCGATGAGGTCTATTTTATGGCAGCAACG-3′. This PCR product was then used as a template for a second PCR using the primer combination 5′-GCTATTTATCCGCTCCATTTGAGCCCGCGAATGAACATATCTGAAGAACC-3′ and 5′-TATAAATTTTTCTATTATCTCATTTTAGTCCTTTAATGTCTCGAGGTATG-3′. In this way, the resultant PCR product from the second reaction now contained the drug-resistance marker flanked upstream by a homology targeting flank corresponding to bp −654 to −596 upstream of the T. brucei cytochrome c gene (gene identification tag Tb927.8.5120) and downstream by a homology flank corresponding to the 23 bases upstream and 33 bases downstream of the stop codon for Tb927.8.5120. Approximately 5 μg of PCR product was used for transfection of procyclic T. brucei and integration into the genome via homologous recombination. For constitutive expression of CxxCH cytochrome c from an ectopic chromosomal locus, this mutant form of T. brucei cytochrome c was PCR-amplified from pKSH035 using the primer combination 5′-CCGAAGCTTATGCCACCAAAGGAGCGTGC-3′ and 5′-ACAGGATCCTTAGTCCTTTAATGTCTCGAGG-3′ (restriction sites introduced into the primers are underlined). The resultant PCR product was digested with HindIII and BamHI and cloned into pDex377  that had also been digested with HindIII and BamHI, creating plasmid pDex377CCres. pDex377CCres was linearized for transfection by digestion with NotI. For disruption of an endogenous cytochrome c allele with a phleomycin-resistance gene, the ORF (open reading frame) for a phleomycin-resistance marker was flanked at its 5′-end by a DNA sequence corresponding to the 251 bp of genomic sequence immediately upstream of the start codon for T. brucei cytochrome c, and at its 3′-end by a DNA sequence corresponding to the 254 bp immediately downstream of the stop codon. Prior to transfection into procyclic T. brucei, this commercially synthesized DNA molecule (purchased from Eurofins) was released from pBluescript (Stratagene) by digestion with NotI. All DNA constructs were sequenced using ABI Prism sequencing technology, and DNA was prepared for transfection by Qiagen Maxi-Prep kits.
For stable transformation, parasites from late exponential-phase cultures (10 ml cultures at ~5×107 cells/ml for C. fasciculata, or ~107 cells/ml for T. brucei) were harvested by centrifugation (800 g for 10 min at 20°C) and re-suspended at 6×107 cells/ml in ZM transfection buffer [132 mM NaCl, 8 mM Na2HPO4, 0.5 mM Mg(CH3COO)2, 0.09 mM Ca(CH3COO)2, pH 7.0] containing 1% (w/v) glucose. Aliquots (0.5 ml) were then mixed with 3–10 μg of DNA in a 0.4 cm electroporation cuvette (Bio-Rad Laboratories) and transformed using an electroporator with three pulses of 1700 V for 0.1 ms with a 0.2 ms rest between pulses. Transformed cells were recovered in 10 ml of fresh culture medium for either 6 h (C. fasciculata) or overnight (T. brucei) before drug selection. Recovered C. fasciculata cultures were diluted 1:100 and selected using hygromycin at 100 μg/ml. Following selection of stable transformants, drug pressure was maintained at 200 μg/ml hygromycin. For T. brucei, rested transformants were inoculated at 1×105 cells/ml in medium containing either blasticidin/HCl (10 μg/ml), blasticidin (10 μg/ml) and hygromycin (50 μg/ml), or blasticidin (10 μg/ml), hygromycin (50 μg/ml) and phleomycin (3 μg/ml). Following the third round of T. brucei transfection to disrupt the remaining endogenous cytochrome c allele, stable transformants were selected in 96-well dishes (200 μl per well) and individual lines from independent transformations were taken for subsequent analysis. Genomic DNA was prepared as described previously . Southern blotting of EcoRI-digested genomic DNA to Hybond-N (GE Healthcare), followed by hybridization against either the T. brucei cytochrome c ORF or coding sequence from the downstream gene Tb9278.5130 (to confirm correct integration of drug-resistance cassettes) was undertaken using an AlkPhos Direct Labelling Kit. This, and detection with CDP-Star (GE Healthcare), were carried out according to the manufacturers’ instructions.
Purification of cytochrome c
T. brucei CxxCH cytochrome c was purified from C. fasciculata using procedures described previously for the purification of WT C. fasciculata cytochrome c . T. brucei CxxCH cytochrome c largely separated from C. fasciculata WT cytochrome c during the first (SP-Sepharose) chromatographic step. The two cytochromes could be distinguished both spectroscopically (see the Results and discussion section) and by SDS/PAGE. Only fractions from the first column overwhelmingly enriched in T. brucei CxxCH cytochrome c were pooled for the second and subsequent chromatographic steps. For small-scale purification of T. brucei CxxCH cytochrome c from C. fasciculata, or from T. brucei itself, a 5 ml Hi-Trap SP-Sepharose column (GE Healthcare) was used. In the case of purification from T. brucei, proteins were batch-eluted from the column using 0.5 M NaCl, rather than by running a gradient. Electrophoretically pure T. brucei CxxCH cytochrome had Rz values of 5.7 (A415, reduced/A280) and 4.7 (A406, oxidized/A280). Note that pNUS-GFPc contains an NdeI site (CATATG) shortly before the KpnI site we used for cloning . The methionine residue from the NdeI site became the start codon for the T. brucei CxxCH cytochrome c expressed in C. fasciculata. Thus the N-terminal cytochrome sequence in our experiments using pure protein was MRGSTMPPKE, where the second methionine residue is the naturally occurring start codon in T. brucei. Note also that, as reported previously , the T. brucei cytochrome c gene which we originally cloned from genomic DNA from 427 strain trypanosomes has an alanine as residue 15 of the mature peptide (using WT T. brucei numbering), whereas in the genome reference strain this residue is valine. The CxxCH cytochrome is an A25C variant of the WT. Absorption spectra were recorded on a PerkinElmer Lambda 2 spectrophotometer. Reduced pyridine haemochrome spectra were recorded using a method described by Bartsch .
For ESI (electrospray ionization)–MS, protein samples were desalted using C4 ZipTips (Millipore) according to the manufacturer's instructions. These samples in 1:1 acetonitrile/water plus 0.1% formic acid were introduced at a flow rate of 10 μl/min by ESI into a Micromass LCT orthogonal acceleration reflecting TOF (time-of-flight) mass spectrometer in positive-ion mode. The mass spectrometer had been calibrated using myoglobin. The resultant m/z spectra were converted into mass spectra using the maximum entropy analysis tool MaxEnt in the MassLynx suite of programs. For MALDI (matrix-assisted laser-desorption ionization)–MS, peptide samples were spotted on to the MALDI target with α-cyano-4-hydroxycinnamic acid. MALDI spectra and MALDI MS/MS (tandem MS) fragment ion spectra were acquired on an AB Sciex 4800 MALDI–TOF/TOF mass spectrometer operated in positive-ion reflectron mode. Data were analysed using 4000 Series Explorer and Data Explorer software. For LC (liquid chromatography)–MS, peptides were loaded into disposable nanospray emitters (New Objective) using a GELoader tip (Eppendorf). Samples were sprayed directly into an Orbitrap mass spectrometer (Thermo). Spectra were acquired at a resolution of 60000 [fwhm (full-width at half-maximum)] over several minutes. Spectra were summed and deconvoluted using Xcalibur and ExtractRaw software (Thermo).
RESULTS AND DISCUSSION
Expression of cytochrome c with a CxxCH haem-binding motif in a trypanosomatid
To look first at expression in trypanosomatids of cytochrome c with a CxxCH haem-binding motif we used C. fasciculata, which is a parasite of insects (non-pathogenic to humans) that grows quickly to high cell densities in culture, and relies on cytochrome-dependent metabolism for viability , making it ideal for biochemical studies. Thus C. fasciculata transformed with an episome  encoding variant T. brucei cytochrome c with a CxxCH haem-binding motif were detergent-extracted to release the proteins. Solubilized cell extracts were loaded on to an SP-Sepharose column to purify the cytochrome c . The column was eluted using a NaCl gradient; red coloured fractions were analysed spectrophotometrically. Two spectroscopically distinct forms of cytochrome c were resolved. Early fractions in the elution peak had, when reduced with dithionite, absorption maxima at ~416, 521 and 550 nm; late fractions had maxima at ~420, 524.5 and 555 nm. The peaks in pyridine haemochrome analysis, which reports on the nature of any covalent haem attachment to protein, were at 550 nm and 553 nm respectively. These results indicate that the C. fasciculata cells produced one form of cytochrome c with two covalent linkages between haem and protein (the early eluting species with its haemochrome maximum at ~550 nm), and another form with one linkage between haem and protein (the late-eluting species, haemochrome ~553 nm). We reasoned that the early-eluting species was likely to be the T. brucei CxxCH variant cytochrome c, and that the late-eluting species was native WT single-cysteine C. fasciculata cytochrome c. The two cytochromes were largely separated by the SP-Sepharose chromatographic step; this was apparent both from spectra and from SDS/PAGE, since the T. brucei variant cytochrome ran larger than WT C. fasciculata cytochrome c. Fractions very significantly enriched in T. brucei CxxCH cytochrome c were pooled and further purified, to electrophoretic homogeneity, by CM-Sepharose and Sephacryl S200 (size-exclusion) chromatography. Absorption and pyridine haemochrome spectra of the pure T. brucei CxxCH holocytochrome c obtained from C. fasciculata mitochondria are shown in Figure 2. The CxxCH variant had absorption characteristics of normal CxxCH cytochrome c, such as those from animals and fungi, i.e. with two thioether bonds between haem and protein, to both cysteine residues of the haem-binding motif.
Absorption spectra of the CxxCH variant of T. brucei cytochrome c purified from C. fasciculata mitochondria
Analysis of the product CxxCH cytochrome c
We assessed in more detail haem attachment to the recombinant T. brucei CxxCH variant cytochrome c. Pure protein was analysed by N-terminal sequencing. This unambiguously confirmed its identity, but it was not possible to sequence as far as the haem-binding site ~30 residues from the N-terminus. The cytochrome was also investigated by ESI–MS. Different batches of purified protein gave masses of 13554.5±1.9 Da and 13552.0±1.4 Da. The calculated (theoretical) mass of the protein including haem (which remains covalently bound during the ESI–MS experiment) is 13453.2 Da. Thus the observed mass of the protein is ~100 Da above that expected. It therefore has additional post-translational modifications (which are common in eukaryotic proteins) and/or haem-attachment side reactions. MS-based peptide mapping of the cytochrome clearly confirmed the presence of a trimethyl-lysine (an extra 42 Da) near the C-terminus (as also found in C. fasciculata cytochrome c [13,21]), but was inconclusive as to the origin of the remaining extra ~58 Da. We therefore isolated and analysed the haem-binding peptide(s) from the T. brucei CxxCH cytochrome c in case any of the extra mass was related to the haem attachment to the protein (which could indicate, for example, haem mis-attachment).
Pure cytochrome was reduced, alkylated and digested with Protease V8 (endoproteinase GluC). The haem-containing peptide fragment was isolated on the basis of its absorbance at both 410 and 215 nm after HPLC of the digest using a C18 column. The peptide was analysed by Edman degradation and found to be a 9-mer with sequence ?AQ?HTGTK. Cysteine gives a blank (‘?’) in the sequencing reaction unless appropriately alkylated , thus ‘?’ is expected for cysteine covalently bound to a haem in a c-type cytochrome. However, from the sequence of T. brucei cytochrome c used in the present study (…GEKLFKGRCAQCHTGTKGGSNG…) the ‘?’ residues are clearly the two cysteine residues. This experiment is therefore further evidence that haem is bound covalently to both cysteine residues of the CxxCH motif. It also shows that the other residues in the haem attachment region are unmodified.
From these Edman sequencing data and absorption spectra (Figure 2) we conclude that the trypanosomatid C. fasciculata can mature cytochrome c with two cysteine residues in the haem-binding motif and that nothing about the IMS redox environment prevents this; the haem becomes attached to both cysteine residues. In bacteria that use the Ccm system for biogenesis of (CxxCH) cytochrome c there is considerable evidence that a disulfide bond between the apocytochrome cysteine residues is an intermediate in the pathway [7,30]. The same may be true in mitochondria that use haem lyase to mature their c-type cytochromes, although the evidence is less clear . One possibility had been that the AxxCH haem-binding motif of euglenozoan mitochondrial cytochromes c evolved to prevent irreversible formation of a disulfide bond within a CxxCH haem-binding motif. The results of the present study show that to be unlikely. It is also of note that maturation in Crithidia of WT (single-cysteine) C. fasciculata cytochrome c alongside CxxCH variant T. brucei cytochrome c indicates that the trypanosomatid cytochrome c biogenesis apparatus is not mechanistically compromised or inhibited by the presence of an apocytochrome with two cysteine residues in the haem-binding motif.
The 9-mer haem-binding peptide isolated from T. brucei CxxCH variant cytochrome c was analysed by high-resolution MALDI–TOF-MS and LC–MS. In MALDI–TOF, masses of 1579.54 and 1579.45±0.05 Da (Figure 3) were obtained. In LC–MS, masses recorded were 1579.55 and 1579.56±0.01 Da. The expected (theoretical) mass of the peptide plus haem plus a proton (to produce the M−H+ ion) is 1565.60 Da. Thus the haem-binding peptide has a mass 13.96 Da higher than expected, implying a further modification, potentially incorrect attachment of haem to protein.
MALDI–TOF spectrum of the haem-binding peptide isolated from T. brucei CxxCH variant cytochrome c purified from C. fasciculata mitochondria
We subjected the haem-binding peptide (the 1579.45 Da ion) to fragmentation by MALDI–MS/MS. Three notable fragment ions were observed, with masses of 616.2, 647.2 and 664.2 Da respectively (each ±0.2 Da) (Figure 3, inset). We assign these as unmodified haem (theoretical mass 616.5 Da), haem plus a sulfur atom minus a proton (theoretical mass 647.6) and haem plus a sulfur atom plus an oxygen atom (theoretical mass 664.6 Da). Before fragmentation, the haem was attached to both cysteine sulfur atoms of the CxxCH haem-binding motif (see above). After fragmentation, three species were observed as the haem was cleaved from the peptide by the breakage of different combinations of carbon–sulfur bonds. Haem itself (616.5 Da) would be produced by the cleavage of both haem-vinyl–cysteine-sulfur bonds. The other two species are both haem with sulfur bound, resulting from the cleavage of one haem-vinyl–cysteine-sulfur bond and one cysteine-β-carbon–cysteine-sulfur bond. (See Figure 1 for an illustration of such bonds in typical cytochromes c.) Observation of both 647.2 Da (haem plus S minus H) and 664.2 Da (haem+S+O) ions indicates that one of the haem-binding cysteine residues has become oxidized and the other has not. The haem itself is unmodified (other than by the attachment via its vinyl groups to protein) when it binds to T. brucei CxxCH cytochrome c, otherwise we could not observe the intact haem 616.2 Da species.
Overall, this detailed mass spectrometric analysis of the haem-binding peptide isolated from T. brucei CxxCH variant cytochrome c indicates a total observed mass of 1579.56±0.01 compared with the expected (calculated) mass of 1565.60 Da for the peptide plus haem M−H+ ion (a difference of 13.96 Da). MALDI–MS/MS reveals oxidation of one haem-binding cysteine sulfur by addition of an oxygen atom (an extra mass of 16.00 Da; recalculated theoretical mass=1581.60 Da). Thus a carbon–carbon bond has also become unsaturated in the haem-binding peptide with the loss of two protons (2.02 Da) (theoretical mass now 1579.58 Da, compared with an observed mass 1579.56). Therefore haem is attached to both cysteine residues of the CxxCH haem-binding motif of the T. brucei cytochrome c, but one of the cysteine residues is oxidized and two protons have been lost to (re)form a carbon–carbon double bond. As shown by Edman degradation, no other residues are modified in the haem-binding peptide. Thus the haem attachment to protein in T. brucei CxxCH variant cytochrome c expressed in Crithidia mitochondria is not the high-fidelity stereospecifically conserved linkage universally observed in natural c-type cytochromes.
These observations are closely analogous to those made for a variant of cytochrome b5 where a cysteine residue was introduced adjacent to a haem-vinyl group. A haem–protein linkage formed, but one product had oxidized cysteine and an unsaturated carbon–carbon bond (, species BII). The MS/MS data obtained for species BII in  were very similar to those reported in the present study (Figure 3). Linkages of this type are characteristic of spontaneous non-enzyme-catalysed reactions between haem and protein cysteine residues [31,32]. The most likely scenario in the present study is that the trypanosomatid cytochrome c biogenesis system recognized and processed the WT naturally-occurring haem-binding cysteine (cxxCH) catalysing its attachment to the haem 4-vinyl group (see Figure 1 for nomenclature), which would leave the artificially introduced Cxxch thiol, unprocessed by the biogenesis machinery, spatially close to the haem 2-vinyl group. A spontaneous (uncatalysed) haem–protein linkage then formed between these functional groups, resulting in an aberrant haem–protein linkage, potentially as a consequence of redox cycling by the haem iron of the cytochrome during respiratory electron transfer . Another possibility is that the haem–protein linkage involving an oxidized cysteine thiol and an unsaturated carbon–carbon bond is the correct product of the as yet unidentified cytochrome c biogenesis machinery of trypanosomatids. However, we consider this unlikely as such a linkage is expected from uncatalysed reactions between haem and cysteine (see above and [31,32]), but would be unique in cytochromes c matured by a biogenesis machinery where the stereospecificity is, to date, invariant. Moreover, the linkage involves a cysteine residue artificially introduced into the cytochrome by site-directed mutagenesis, i.e. which the trypanosome cytochrome c biogenesis machinery would not normally be exposed to. Finally, but significantly, the CxxCH cytochrome with mis-attached haem results in less-efficient respiration in trypanosome cells (see below).
Is CxxCH cytochrome c functional in the trypanosome respiratory chain?
Although our experiments in C. fasciculata showed that incorrectly matured CxxCH cytochrome c could accumulate in cells, they did not indicate whether cytochrome c function was compromised (since WT cytochrome c was also present). To address the issue of function, we performed gene-replacement studies in the procyclic forms (tsetse fly mid-gut stage) of another trypanosomatid T. brucei. At the time we began our study a complete and well-annotated genome sequence was available for T. brucei but not for C. fasciculata. Coupled to the wider range of tractable gene expression systems available for African trypanosomes, this meant it was more straightforward to design a gene-replacement strategy for T. brucei than Crithidia. Importantly, all of the available data [11,14,15,21] suggests that these trypanosomatid species are equivalent in terms of their cytochrome c and cytochrome c maturation. However, owing to the considerably lower cell densities seen in late exponential/early stationary phase cultures than for C. fasciculata, it would have been impractical to isolate sufficient protein from T. brucei for the biochemical experiments described above.
Numerous studies indicate that cytochrome-c-dependent respiration is essential for cellular viability of procyclic T. brucei (e.g. [33,34]), and from the genome sequence it was straightforward to identify intergenic sequence 5′ and 3′ to the cytochrome c ORF, which facilitated gene replacement via homologous recombination. Trypanosomes are diploid. Thus we sequentially replaced one cytochrome c gene from T. brucei with a blasticidin deaminase drug-resistance marker (generating a CYTC+/− cell line), introduced an ectopic copy of CxxCH cytochrome c that was expressed constitutively from another chromosomal location (generating CYTC+/−/c−c), and finally replaced the second endogenous copy of the WT cytochrome c gene with a phleomycin-resistance marker (generating CYTC−/−/c−c). All transfections yielded viable transformants within standard time-windows (10–14 days for populations placed under drug selection; 20–23 days for clonal transformants selected in 96-well plates). Genomic DNA isolated from transformed cell lines was used in Southern blotting to confirm targeted integration of drug-resistance cassettes and the loss of WT cytochrome c genes from the genome of CYTC−/−/c−c cells (Supplementary Figure S1 at http://www.BiochemJ.org/bj/448/bj4480253add.htm). DNA encoding the ectopic cytochrome c gene present in the CYTC−/−/c−c mutant was also isolated by PCR amplification and sequenced, confirming the exclusive presence of the desired A25C mutation (i.e. CxxCH cytochrome c).
Thus, as cytochrome c is essential, the introduced CxxCH variant cytochrome c was functional in respiration in T. brucei, at least in culture, when WT cytochrome c was absent. CxxCH-variant cytochrome was partially purified (at a small scale) from CYTC−/−/c−c cells by SP-Sepharose chromatography to establish the nature of the haem attachment. The reduced cytochrome had its absorption maxima at 550 nm and pyridine haemochrome α-band maximum at 550.5 nm (Figure 4). In contrast, cytochrome c from WT T. brucei had these maxima at 553–556 nm (broad peak) and 553.0 nm respectively (Figure 4). Therefore, as we found to be the case for the CxxCH variant T. brucei cytochrome c purified from C. fasciculata mitochondria, in the cytochrome purified from T. brucei itself the haem is attached through two haem–protein linkages.
Absorption spectra of the CxxCH variant of T. brucei cytochrome c purified from T. brucei mitochondria
To look further at the use of CxxCH variant cytochrome c we investigated CYTC−/−/c−c cells, and the heterozygote and parental strain from which these mutants were derived, following incubation with the mitochondrial respiratory chain inhibitors antimycin A and sodium azide (Figure 5). These inhibitors inhibit Complexes III and IV respectively, which both interact with cytochrome c in the electron transport chain. The effect of each inhibitor on the parental 427 cells was consistent with data available in the literature (e.g. [21,34]), and in duplicate experiments the CYTC−/−/c−c mutant appeared to show slightly increased sensitivity towards both inhibitors. The CYTC−/−/c−c cells were notably more susceptible than the parental or heterozygote strains to the addition of antimycin A at 0.25 μg/ml (Figure 5B). Since overall expression levels of cytochrome c were comparable between mutant and WT trypanosome cultures (see below), this suggests that replacement of endogenous T. brucei cytochrome c with the CxxCH variant carries a fitness cost.
The effect of mitochondrial respiratory chain inhibitors on the growth of CYTC−/−/c−cT. brucei which contains only CxxCH variant cytochrome c
The amount of holocytochrome c produced in trypanosomatid mitochondria is limited
When either WT T. brucei cytochrome c or the CxxCH variant was expressed from an episome in WT C. fasciculata (i.e. not in a cytochrome c deletion strain), and when the CxxCH cytochrome c was subject to potential overexpression from an ectopic chromosomal location and off a strong transcription promoter in T. brucei itself, the total amount of holocytochrome c produced by cells was similar to the amount of cytochrome produced by the respective WT/parental strains. For example, assuming roughly equal absorption coefficients for each form of cytochrome c, the total amount of cytochrome c produced per unit of cells of T. brucei was the same for both the WT cells and the CxxCH variant expressing (CYTC−/−/c−c) cells (e.g. Figure 4). Similarly, the total amount of cytochrome c produced was the same per unit of cells in untransformed C. fasciculata cultures, and in Crithidia expressing their endogenous cytochrome c plus either T. brucei CxxCH cytochrome c or WT T. brucei cytochrome c. Thus it appears that the potential for holocytochrome c production in trypanosomatids is limited. This may well be due to the capacity of the trypanosomatids’ unique cytochrome c biogenesis machinery. Alternatively, since Crithidia and trypanosomes are haem autotrophs , i.e. they cannot synthesize haem and must acquire it from an external source, cytochrome c accumulation is plausibly regulated by the rate of haem uptake into the cell or its delivery across the cytosol to the mitochondrion. Note that, in our experiments, haem was abundant in the culture medium (without which the cells could not grow); thus haem availability outside the cell was not the limiting factor for cytochrome synthesis.
Evolutionary considerations indicate that ancestrally in Euglenozoa and in the closest related phyla with aerobic mitochondria, Heterolobosea and Jakobida, typical CxxCH-type cytochromes c and c1, and the Ccm system for cytochrome c biogenesis, were all present . However, before the divergence of the various euglenozoan species from their last common ancestor, the Ccm system was lost. It was not replaced by haem lyase as occurred in all other eukaryotes where the Ccm system is absent, but cytochromes c and c1 remain , but by a novel biogenesis machinery. The results of the present study indicate that faulty maturation of CxxCH cytochrome c can be tolerated in euglenozoan cells (e.g. Figures 4 and 5, and Supplementary Figure S1), at least in the short term, which would permit degeneracy followed by loss of the multi-component Ccm system (and its replacement by the new biogenesis machinery), before the single-cysteine cytochromes c evolved.
At first glance, our observation that cytochrome c with a CxxCH haem-binding motif is functional in trypanosome mitochondria (Figure 4 and Supplementary Figure S1) suggests that this protein transfers electrons comparably with the WT (AxxCH) cytochrome. A corollary is that their reduction potentials are not significantly different, consistent with biophysical studies which indicate that the midpoint potentials of cytochromes c differ by no more than 20 mV whether the haem is attached through one or two thioether bonds [12,36]. However, at the point in euglenozoan evolution where the Ccm system (which cannot attach haem at an appreciable rate to single cysteine haem-binding motifs [21,37]), was lost, the CxxCH haem attachment motif presumably was disadvantageous, since the N-terminal cysteine residue is absent from the haem-binding motif of mitochondrial cytochromes c in all extant euglenozoans analysed . Consistent with this assertion, our experiments with respiratory chain inhibitors (Figure 5) point towards some fitness cost associated with the use of CxxCH cytochrome c in an extant trypanosomatid. A fitness cost of this kind could have provided a selective pressure in an early euglenozoan (which had the novel cytochrome c biogenesis machinery) that led to loss of the now improperly matured CxxCH cytochrome c and evolution of the xxxCH form. The non-compatibility of a newly acquired or newly evolved euglenozoan cytochrome c biogenesis machinery with correct maturation of a CxxCH haem-binding site, and the need for compensatory mutations to preserve function, may have been even more acute for cytochrome c1 where phenylalanine and proline are unexpected residues conserved across several hundred million years of euglenozoan evolution in an FAPCH haem-binding site [14,38]. Overall, the available data now suggest that evolution of the single-cysteine cytochromes c in euglenozoans was driven by the properties of the cytochrome c biogenesis machinery these organisms acquired or evolved to replace the Ccm system, rather than by other factors such as the environment of the mitochondrial IMS, or the relative biophysical properties of (correctly matured) single- and double-cysteine cytochromes c.
Michael Ginger and James Allen designed the experiments. All authors carried out experiments. Michael Ginger and James Allen analysed data and wrote the paper.
We thank Elena Nomerotskaia for technical assistance; David Staunton (Department of Biochemistry, University of Oxford, Oxford, U.K.), Benjamin Thomas and Gabriela Ridlova (Dunn School of Pathology, University of Oxford, Oxford, U.K.) for MS; Antony Willis (Department of Biochemistry, University of Oxford, Oxford, U.K.) and Jeff Keen (Leeds University, Leeds, U.K.) for Edman sequencing; and Stuart Ferguson for helpful discussions.
This work was supported by the Biotechnology and Biological Sciences Research Council [grant number BB/D019753/1 (to J.W.A.A.)] and The Royal Society (to M.L.G.). J.W.A.A. was a Biotechnology and Biological Sciences Research Council David Phillips Fellow, M.L.G. was a Royal Society University Research Fellow and K.A.S. was the W.R. Miller Junior Research Fellow, St Edmund Hall, Oxford.