Purification and characterization of cytochrome c oxidase from the insect trypanosomatid Crithidia fasciculata

ELSEVIER

Molecular and Biochemical Parasitology 79 (1996) 47-59

Purification and characterization of cytochrome the insect trypanosomatid Crithidia

c oxidase from

Dave Speijer”, Anton 0. Muijsers”, Henk Dekkerb, Annett de Haan”, Cornelis K.D. Breek”, Simon P.J. Albrachtb, Rob Benne”.* “E.C. Slater Institute, University of’ Amsterdam, Academic Medical Centre, Meibergdrecf I.5, 1105 AZ, Amsterdam, The Netherlands bE.C. Slater Institute, University oj’ Amsterdam, Plantage Muidergracht 12, 1018 TV, Amsterdam. The Netherlands Received 26 January 1996; revised 10 April 1996; accepted 29 April 1996

Abstract Cytochrome c oxidase was purified from the mitochondrial lysate of the insect trypanosomatid Crithidia fasciculata with the aid of a methyl hydrophobic interaction column in a rapid one-step procedure. The purified complex displayed all characteristics expected from a eukaryotic cytochrome c oxidase: the presence of Cu, in electron paramagnetic resonance analysis, a characteristic 605 nm peak in reduced-minus-oxidized optical spectroscopy, and the capacity to efficiently oxidize homologous, but not heterologous, cytochrome c. Two-dimensional PAGE showed that C. ,f&ciculata cytochrome c oxidase consists of at least 10 different subunits. N-terminal sequences were obtained from the six smallest subunits of the complex, one of them showing significant similarity to Neurospora crassa

cytochrome c oxidase subunit V. The N-terminus Keywords:

RNA editing; Trypanosomes;

of each of the four largest subunits was found to be blocked.

Mitochondrion;

Cytochrome

c oxidase; 2D gel analysis; Kinetoplast

1. Introduction

Abbreviations: cox, cytochrome c oxidase; HIC, hydrophobit interaction chromatography; mt, mitochondrion, mitochondrial; nt, nucleotide(s). * Corresponding author. Tel: + 31 20 5665159; fax: + 31 20 6915519.

0166-6851/96/$15.00

0 1996 Elsevier

PII S166-685 1(96)02648-5

Science

B.V. All rights

reserved

In trypanosomatid protozoa and other species belonging to the order kinetoplastida, mt RNAs are post-transcriptionally edited by insertion and deletion of uridylate residues (for reviews, see [l-5]). The information for this remarkable process is provided by small (55-70 nt) guide (g)

48

D. Speijer et al. 1 Molecular and Biochemical Parasitology 79 (1996) 47-59

RNAs, which are complementary to edited sequences including G-U basepairing [6]. The extent to which mt RNAs are edited varies between transcripts in a species-dependent fashion, but in virtually all cases editing is essential for the production of translatable RNAs. For example, in all trypanosomatids studied, cytochrome c oxidase (cox) subunit 2 mRNA is edited by the insertion of 4 Us at an internal position, which removes a genomically encoded frameshift. In Crithidiu fusciculata and Leishmania tarentolae, editing of the 5’ region of cox 3 mRNA creates an in-frame AUG codon which is lacking in the unedited sequence. Editing of cox 3 mRNA in Trypanosoma brucei, however, is much more spectacular, the size of the transcript being doubled by the insertion and deletion of 547 and 41 Us, respectively [7,8]. This way of extensive editing is found in nine of the 17 or 18 mt mRNAs in T. brucei [ 1,4]. The reason for the existence of RNA editing is far from clear. It has been suggested that it would provide the trypanosomes with an extra level of regulation of mt gene expression (see reviews). In addition, it could be envisaged that RNAs are alternatively edited (analogous to alternative splicing), allowing the production of multiple proteins from one gene. Although alternatively edited RNAs have indeed been observed in T. brucei [ 1,4], their relevance is unclear and proteinsequence data that could shed further light on this hypothesis are lacking. This even implies that as yet there is no formal proof that edited mRNAs are indeed used by the trypanosomal mt proteinsynthesizing machinery, the only evidence being provided by an experiment in which a weak signal was obtained by probing Western blots of a cytochrome-enriched fraction from L. turentolae with an antiserum generated against the 11 carboxy-terminal amino acids encoded by edited cox 2 mRNA [9]. As part of a project to analyse mt proteins encoded by edited mRNAs in kinetoplastids, we have purified and characterized cytochrome c oxidase from C, fasciculata. All eukaryotic cytochrome c oxidases studied thus far are multi-component complexes composed of three large mitochondrially encoded subunits (cox 1, 2,

and 3), which form the catalytic core of the enzyme, and up to 10 smaller subunits encoded by the nuclear genome (reviewed in [lo,1 11). Cytochrome c oxidases contain four prosthetic groups: two hemes (a and a_?, both located in subunit 1) and two copper centres, Cu, (two Cu atoms, located in subunit 2) and Cu, (one Cu atom, located in subunit l), [12]. Heme a and Cu, are found exclusively in cytochrome c (and not in other) oxidases [13]. Therefore, we have used the occurrence of the specific heme a-derived absorption peak at 605 nm in a reduced-minus-oxidized spectrum as a diagnostic hallmark to monitor the purification of cytochrome c oxidase from C. fasciculata, together with its capacity to oxidize cytochrome c. The enzyme proved to be difficult to purify because of an unusually tight interaction with the other complexes of the respiratory chain. Nevertheless, here we report the development of a purification procedure with which a substantial part of the enzyme could be isolated without appreciable cross-contamination. We estimate this preparation to be 80% pure. We further established that cytochrome c oxidase from C. fasciculata indeed has a Cu, site and is a large multi-component complex.

2. Materials and methods 2.1. Cell growth and preparation of mt fractions

C. fasciculata was grown (with shaking and aeration) in batches of 10 1, as described in [14], to a density of approximately 1.1 x lo* cells ml - ’ . Mt vesicles from C. fasciculata were isolated according to the method described in [15]. Routinely, the mt vesicle preparation was enriched 50to lOO-fold, as judged by Northern-blot analysis with a mt DNA segment containing the rRNA genes as a probe. 2.2. Pur$cation

of cytochrome c oxidase

Mt vesicles from ml O.l-0.5% (v/v) the experiment) in pH 7.5 (or 20 mM

10 1 of cells were lysed in 15 Triton X-100 (depending on 50 mM potassium phosphate, Hepes-KOH, pH 7.6) and 1

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

p M phenylmethylsulfonyl fluoride, followed by centrifugation at 10 000 x g for 15 min at 4°C. The pellets were extracted with 3% (w/v) lauryl maltoside in the presence of 300 mM KCl. After centrifugation at 10 000 x g for 15 min at 4°C to remove insoluble debris, the resulting supernatant (10 ml), containing approximately 50 mg protein as determined by the method of Lowry [16], was loaded on a 20-ml methyl-HIC column (Biorad), equilibrated in 1.5 M (NH&SO, in buffer A (50 mM potassium phosphate, pH 7.5 and 0.05% (w/v) lauryl maltoside). The column was eluted with 850 mM (NH&SO, in buffer A to obtain a fraction containing all the respiratory chain complexes. Following the 850 mM (NH&SO, step, the methyl-HIC column was eluted with buffer A in the absence of (NH&SO,. This resulted in a fraction containing the purified cytochrome c oxidase. The chromatographic behaviour of cytochrome c oxidase from C. fusciculata on methyl-HIC columns was established by trial runs with 1 mg of protein on a 5-ml methyl-HIC column and a 1.5-O M (NH&SO, gradient in buffer A. In order to establish purity, gel filtration analysis was performed by loading 1 mg of protein (in 1 ml of the 0 M (NH&SO, fraction) on a 25-ml Superose 6 column (Pharmacia) equilibrated in buffer A. 2.3. Electron

Paramugnetic

Resonance

(EPR)

EPR measurements at the X-band (9 GHz) were obtained with a Bruker ECS 106 EPR spectrometer with a field-modulation frequency of 100 kHz. Cooling of the sample was performed with an Oxford Instruments ESR 900 cryostat equipped with ITC4 temperature control. The magnetic field was calibrated with an AEG Magnetic Field Meter. The X-band frequency was measured with an HP 5350B microwave frequency counter. 2.4. PAGE analysis 2D PAGE analysis was performed as described in [17] with a non-denaturing gradient (5- 13% (w/v) polyacrylamide) or linear (5% (w/v) or 6% (w/v) polyacrylamide) ‘blue native’ gel in the first

49

dimension followed by a denaturing Tris-Tricine/ SDS gel (16% (w/v) polyacrylamide) in the second dimension. Approximately 250 pg protein/lane was layered onto a first dimension minigel in 1.1 M aminocaproic acid, 60 mM Bistris, 0.5-3.00/o laurylmaltoside (depending on the experiment) and 0.4% Serva blue G. After running the first dimension for 4 h at 60 V as described in [ 171, which results in the separation of intact, enzymatically active respiratory chain complexes, the separate lanes were excised, incubated for 1 h in 1%) SDS, 1% P-mercaptoethanol and put on top of a second dimension Tris-Tricine/SDS gel [ 171. The second dimension was run for 4 h at 100 V. Tris-Tricine/SDS gels were also used for conventional 1D analysis of the subunit composition of purified material. Staining was done with Coomassie brilliant blue. In order to establish the identity of the different complexes separated in the first dimension gel, gel slices (100 ~1) were incubated for 3 h at 4°C in 500 ~1 50 mM Tricine, 15 mM Bistris pH 7 and 0.1% laurylmaltoside to elute the complexes. ATPase activity (complex V) was measured essentially as described in [18] in the presence or absence of oligomycin, with a 2-min preincubation step with 1% lecithin. Activity of the bc, complex was measured by following the reduction of cytochrome c at 550 nm by Q2H2 (synthesized by A.F. Hartog), in the presence or absence of antimycin as described in [19]. Cox activity was measured with horse cytochrome c as a substrate as described in Section 2.5. A conventional NADH-dehydrogenase appears to be absent in cultured C. jirscicul&a, in agreement with [20]. 2.5. Spectrophotometric measurements

and uctivit? of cytochrome c osiduse

Visible spectra were obtained with the aid of a Beckman DU-70 spectrophotometer. Cuvettes (60 ~1) were filled with the samples to be tested and scanned from 400 to 700 nm, giving the absolute spectrum of the oxidized state. If necessary, the sample was treated with K,Fe(CN), to oxidize it completely. After administration of a trace amount of sodium dithionite, the scan was repeated giving the absolute spectrum of the re-

50

D. Speijer et al. 1 Molecular and Biochemical Parasitology 79 (1996) 47-59

duced state. For the reduced-minus-oxidized spectra the scan of the oxidized state was stored as zero value. Cox-containing fractions (10 /II per assay in a total volume of 400 ~1) were tested for the capacity to oxidize reduced horse heart cytochrome c from Sigma (30 PM in 30 mM potassium phosphate, pH 7.4, 0.1% lauryl maltoside and 0.5 mM antimycin) by following the disappearance of 550nm absorption in the presence or absence of KCN.

calculation was based on the absorption coefficient (reduced-minus-oxidized) of 24 mM - ’ . cm-’ at 605 nm of bovine cytochrome c oxidase [21]). Only the initial linear decrease in absorbance (first minute) was used for activity calculations. Values are the means of two experiments. Cytochrome c from horse and cow are from Sigma, cytochrome c from C. fasciculata was obtained by the method of Margoliash and Walasek [22] after cell rupture by hypotonic lysis according to [15]. Yeast, human heart and rat heart cytochrome c were also isolated according to [22].

2.6. Steady-state kinetics with purified cytochrome c oxidase

2.7. N-terminal sequencing

Cytochrome c from different sources, as indicated in Table 1, was reduced with the aid of a trace amount of sodium dithionite. Sodium dithionite was removed with a G-25 gel filtration column and oxidation was followed spectroscopically by monitoring the disappearance of absorption at 550 nm (or 555 nm for C. fasciculata cytochrome c). All measurements were done in 30 mM potassium phosphate (pH 7.4) and 0.1% lauryl maltoside in a total volume of 600 ~1. Purified cytochrome c oxidase (approximately 5 mg after removal of contaminants by a Pharmacia SMART Superose 6 column) showing only the 605-nm absorption peak upon reduction, was used immediately after isolation and diluted 500fold to a concentration of about 10 nM (the

PAGE was followed by semi-dry blotting onto polyvinylidene fluoride membranes (Biorad), according to the protocol provided by the manufacturer. N-terminal sequencing was done with a Perkin Elmer/Applied Biosystems Precise 494 protein sequencer, or a Beckman-Porton LF 3200 protein sequencer.

Table 1 Steady-state kinetics of C. fasciculata cytochrome c oxidase Cytochrome c from

PM 2

C. fusciculata Man Yeast Rat Horse cow

5

>I40

111 3.29

ND ND ND ND

10

6.16

1.91 1.79 1.25 0.76

110 6.76

2.14 1.61 ND ND

Values given are the turnover numbers (SK’) obtained with the indicated concentrations of reduced cytochrome c of the different organisms, in the presence of approximately 10 nM of cytochrome c oxidase from C. ,fasciculata (see Section 2); ND. not determined.

. 3. Results 3.1. PuriJication of cytochrome c oxidase from C. fasciculata

Purification procedures of cytochrome c oxidase from such diverse eukaryotic organisms as cows [23 -251, Dictyostelium discoideum [26] and Saccharomyces cerevisiae [27,28] make use of differential extraction of mt membranes with nonionic detergents [23,25,28], of hydrophobic interaction chromatography with phenyl or octyl Sepharose columns [27] and of anion-exchange chromatography [25,28]. With mt membranes of C. fasciculata, however, the use of varying concentrations of detergents such as Triton X-100, lauryl maltoside or deoxycholate in differential extraction procedures did not result in the separation of the oxidase from the other complexes of the respiratory chain, as judged from cox-activity measurements and from spectrophotometric and 2D-PAGE analyses (results not shown, see Section 2). Moreover, C. fasciculata cox could not be eluted from alkyl and phenyl hydrophobic inter-

51

D. Speijer et al. 1 Molecular and Biochemical Parasitology 79 (1996) 47-59

Table 2 Purification of cytochrome c oxidase from C. fhsciculata Purification step

Amount of protein (mg)

Specific Activity (nmol cytochrome c oxidized min-’ mg- ‘)

Enrichment (fold)

mt vesicles mt membranes 850 mM (NH&SO, fraction methyl HIC 0 mM (NH&SO, fraction methyl HIC

100 12.8 8.8

1.9 19 29

I” IO I5

33

17

2.9

Cytochrome c oxidase was purified from mt vesicles prepared from 10” C. fasciculata cells (as described in Section 2). “Cytochrome c oxidase activity in whole-cell extracts was below detection levels. The mt vesicle preparation was enriched 75-fold, as judged from Northern blot analyses of mt and total RNA from C. fasciculatu.

action or anion-exchange columns (e.g. Pharmacia Mono-Q). Both features may be related to the extreme hydrophobicity of the complex (compare for example the inferred amino acid sequences of trypanosomatid cytochrome c oxidase subunits 1, 2 and 3; see [7,29-321). We therefore developed a new purification procedure for C. fasciculata cox. After solubilization of the mitochondrial vesicles with 0.1% (v/v) Triton X-100, practically all the hydrophilic macromolecules were present in the supernatant, whereas only trace amounts of the respiratorychain complexes were present. This was checked by Northern-blot analysis of mt rRNAs as representatives of hydrophillic macromolecules and by spectroscopic analysis, cox-activity measurements and 2D PAGE for the respiratory-chain complexes (not shown, see below). After this, the pellet was extracted twice with 3% (w/v) laurylmaltoside in the presence of 300 mM KC1 and the combined extracts, containing all cox activity, were loaded onto a methyl-HIC hydrophobic interaction column. When a 1.5-O M gradient of (NH&SO, was applied to this column, cox activity reproducibly eluted in two peaks at 1.1 M and 0.7 M (NH&SO,, respectively. Since the two peaks together contained all proteins that eluted from the column, we routinely used stepwise elution with 850 mM (NH&SO, and without added salt to separate the two fractions. The procedure and enrichment are summarized in Table 2. This table shows that the 850 mM (NH&SO, fractions contained 70% and the 0 mM (NH&SO, frac-

tions 30%, respectively, of catalytically active cox (see Section 4). The protein content of these fractions was analysed by 2D PAGE analysis consist(‘blue native’) ing of a non-denaturing first-dimension gel, which separates complete, enzymatically active respiratory-chain complexes [ 171, followed by a denaturing second-dimension gel. As shown in Fig. lB, the 850 mM (NH&SO, fraction contained numerous complexes, all of which were also present in the complete mt membrane fraction (Fig. 1A). Activity measurements of complexes A’ and A following elution from the first dimension gel and N-terminal sequence analysis of proteins present in complexes suggests that they represent, most likely, dimeric and monomeric forms of C. fasciculata ATP synthase (results not shown, see legend to Fig. 1). Similar analyses identified complex C as the C. fasciculata bc, complex, containing cytochrome c, and four other known C. fasciculata bc, proteins [33] (results not shown), whereas we could not identify complex D. In the 0 mM (NH&SO, fraction one complex (designated B’) predominated (Fig. 2). Since this complex displayed cyanide-sensitive cox activity following its elution from the non-denaturing first-dimension gel, we concluded that it represents C. fasciculata cox. This was confirmed by measurements of the spectral and catalytic properties of the 0 mM (NH&SO, preparation (see below). From 2D PAGE (Fig. 2) and Superose 6 gel filtration analysis (not shown) we estimated that in the 0 mM (NH,),SO, fraction, complex B’ was approximately 80% pure and

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

52

kDa

A’AB’C

D

A’

A

B’C

DB

kDa

Fig. 1. 2D gel analysis of mt membrane proteins of C. fasciculata. A. Thirty micrograms of protein of a mt vesicle preparation lysed with 0.5% (v/v) Triton X-100. B. Twenty micrograms of protein of a 850 mM (NH&SO, fraction from the methyl-HIC column (see Section 2). 2D gel analyses were performed as described in Section 2. In the first dimension, blue native gels of either 5% (w/v) (panel B) or 6% (w/v) (panel A) were used. In the second dimension 16% (w/v) Tris/Tricine gels were used. A. A’ indicates ATP synthase; B, B’, cytochrome c oxidase; C, bc, complex (cytochrome c reductase); D, unknown complex. Activity assays of eluted material are described in Section 2. The 2D was calibrated with molecular-weight markers, the size of which is indicated. The beginning of the arrow above the figure coincides with the top of the gel slice taken from the first dimension. Staining was done with Coomassie brilliant blue.

consisted of approximately 10 different subunits (see below). A comparison of its subunit composition with that of bovine cytochrome c oxidase is shown in Fig. 3. In the remainder of this paper we refer to complex B’ as C. fusciculata cox. 3.2. Spectral properties of cytochrome c oxiduse from C. fasciculata To characterize the purified cox in the 0 mM (NH&-SO, fraction and to determine whether or not other respiratory-chain complexes were present as contaminants, its spectral properties were checked in two ways. First, an absorption scan was made with visible light from 400-700 nm. This was done with the complex in the oxidized state, but also following reduction with a trace amount of dithionite. Fig. 4A shows the two absolute spectra, while Fig. 4B shows the reducedminus-oxidized spectrum. In Fig. 4A as well as in Fig. 4B the shift in the Soret peak from approximately 420 to approximately 446 nm (indicative of cytochrome c oxidase) can be clearly seen. The 605-nm peak (most pronounced in Fig. 4B) is a clear indication of the presence of cytochrome c oxidase, while the virtual absence of an absorp-

tion peak at 560 nm indicates the absence of the of the preparation expressed as A 440JA 420red is 1.28, so most of the preparation had lost its native form (compare for example [34], and see Section 4). The second kind of spectral analysis was performed with EPR. With this approach most of the metal ions from the catalytic centers of the respiratory chain can be specifically recognized. In the case of cytochrome c oxidase the signal of the Cu, centre is diagnostic. EPR is a relatively insensitive method (Cu, concentrations should be above approximately 5 ,uM). EPR measurements with the 850 mM (NH&SO, fraction routinely gave a very clear Cu, signal (results not shown). The 0 mM (NH,),S04 fractions, however, were generally too diluted to detect EPR signals of Cu, and concentration procedures invariably led to denaturation and loss of signal. Only when relatively large amounts of the 3% (v/v) lauryl maltoside extract were loaded onto the methyl-HIC column, were we able to generate a 0 mM (NH&SO, fraction concentrated enough to obtain a convincing Cu, signal (Fig. 5). The EPR spectrum of the untreated oxidase is displayed in Fig. 5, trace B. Several features typical for cybc, complex. The reducibility

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

tochrome c oxidase can be recognized. The main part of the spectrum is caused by the S = l/2 system of the Cu, centre [35], giving a nearly axial signal with g, = 2.184 and g,,, values around 2.0 [36]. The lines are rather broad due to interaction of the unpaired electron with the two Cu nuclei, both having a nuclear spin of 1.5 [35]. This interaction also leads to the typical features (in Xband) around apparent g values of 2.072 (top), 1.987 (trough) and a shoulder at 1.963. At g = 2.255 the top of the g, line (at 2.233) of the low-spin heme a can also be observed. Upon reduction with dithionite, these signals completely disappeared and the spectrum in trace A was obtained. The major part of the reduced spectrum can be interpreted as due to the reduced [2Fe-2S] cluster of succinate dehydrogenase (gxyZ = 1.921, 1.934, 2.025) [37]. The radical signal at g = 2.003 is probably due to partially reduced flavin in this enzyme. The signal typical for the reduced Rieske [2Fe-2S] cluster (gxyZ = 1.787, 1.892, 2.024) [38] was absent in this preparation. b

A

8’C

kDa

53

kDa 94 67

43

-I

30

- II

1 2 43 : 7

- III 20.1

8 9

14.4

10

- IV _Va :Vb Via : ;I,” _ Vlla I Vllblc VIII

Fig. 3. Comparison of cytochrome c oxidase from C. fasciculata and cow. Purified cytochrome c oxidase from C. fbciculara (2.5 pg) (lane 1) and 7.5 pg of bovine cytochrome c oxidase (lane 2) were analysed on a 16% (w/v) Tris/Tricine gel (see Section 2). Bovine cox subunits are indicated on the right (I-VIII, nomenclature according to [23]), molecular weights of marker proteins are given in the middle and numbers of C. fasciculata bands are depicted on the left. Staining was done with Coomassie brilliant blue.

Fig. 2. 2D gel analysis of purified cytochrome c oxidase from C. ,fasciculata; 7.5 mg of protein of the 0 mM (NH&SO, fraction from the methyl-HIC column (see Section 2 and Table 1) was loaded. 2D gel analyses was performed as described in Section 2. In the first dimension, a blue native gel of 5% (w/v) was used. In the second dimension a 16% (w/v) Tris/Tricine gel [17] was used. For A, B’ and C: see legend to Fig. 1. The subunits of cytochrome c oxidase are indicated on the left of the panel, while the molecular-weight markers are indicated on the right. The beginning of the arrow above the figure coincides with the top of the gel slice taken from the first dimension. Staining was done with Coomassie brilliant blue.

However, visible spectrophotometric analysis performed as in the experiment of Fig. 4 and 2D PAGE analyses showed a small amount of the bc, complex to be still present. Nevertheless, we interpret the EPR results to support our conclusion that complex B’ of Figs. 1 and 2 is C. fusciculata cox.

3.3. Catalytic properties of cytochrome c oxidase from C. fasciculata The purified

cytochrome

c oxidase from C. activity assays. The results, given as the turnover number (SK’) of cyto c h rome c oxidized per molecule of

fusciculata was also tested in steady-state

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

54

cox, are shown in Table 1. The purified fraction was evidently capable of oxidizing reduced cytochrome c from different sources. It shows, however, a very clear preference for its natural substrate: e.g. the oxidation of bovine cytochrome c proceeded 150 times slower than that of endogenous cytochrome c. C. .fasciculata cytochrome c has a number of distinguishing features that could be causally related to this preference. First, the heme group is linked to the protein via a single thio-ether linkage, resulting in a shift of the &,,, from 550 to 555 nm. In addition, the redox potential is about 25 mV higher, whereas the isoelectric point is lower (8.8 instead of 10.0 [39]). Last but

0.5

i?

3.4. Subunit composition of C. fasciculata cytochrome c oxidase and peptide microsequence analysis

A

1

2

not least, trypanosomatid cytochrome c appears to be more hydrophobic than the other cytochrome c’s (compare for example the.sequences in [40]). This characteristic could make it a better substrate for endogenous cox 2 (the subunit containing the binding site for cytochrome c), which is much more hydrophobic than all cox 2’s analysed thus far. In line with this, human cytochrome c, which of the other cytochrome c’s is the most hydrophobic [41,42], proved to be a significantly better substrate than cytochrome c from either yeast, rat or cow.

0.025

lB

-0.023

-0.035 I 400.0

I 460.0

I 520.0

I 580.0

I 640.0

I 700.0 "Ki

Fig. 4. Spectral analysis of C. jhsciculata cytochrome c oxidase. On the horizontal axis the wavelength is indicated in nm and on the vertical axis the absorbance. Cox (approximately 1.5 ,uM) from the 0 mM (NH&SO4 fraction, was spectrally analysed as described in Section 2. The black line in A represents the oxidized spectrum (relative to buffer) obtained with K,Fe(CN),. The grey line shows the reduced spectrum obtained with sodium dithionite. In B, the oxidized spectrum is subtracted from the reduced spectrum.

Analysis of purified C. fasciculata cox by 1D and 2D PAGE allowed assessment of its subunit composition. Approximately 10 polypeptides migrated as a large multi-component complex in the non-denaturing first-dimension gel of Fig. 2 (numbered 1- 10, see also Section 4). As mentioned above, complex B’ was capable of cyanide-sensitive cytochrome c oxidation, as measured in the eluate of relevant gel slices. The same polypeptides showed up as major bands when purified C. ,fasciculata cox was directly analysed by SDS PAGE (Fig. 3) and they were also present in complexes B and B’ in the 850 mM (NH&SO, fraction (Fig. 1B) and in the crude extract (Fig. 1A). Comparison of Fig. lB, Figs. 2 and 3 reveals that the second and third largest subunits are difficult to separate in the second-dimension SDS gels, particularly when high amounts of protein are used. This behaviour is also observed for cox 2 and 3 from other organisms (e.g. man [43]). We have undertaken sequence analysis of the N-terminus of all proteins that are present in the cox complex at different stages of the purification procedure, following both 1D and 2D PAGE. Sequences could be generated for six of the smallest subunits (subunits 5-10, Table 3). A low but significant similarity was found between the N-terminal sequence of C. fasciculata subunit 8 and that of subunit V of Neurospora crassa cytochrome c oxidase (the mammalian subunit IV homologue [44]). The full-length cDNA sequence

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

2.4

2.3

2.2

2.1

2.0 G -

1.9

1.6

1.7

VALUE

Fig. 5. EPR spectra at 35 K of C. jhsciculata cytochrome c oxidase. Concentrated cox (5 pm) was obtained as described 2 and Section 3. A (upper spectrum): cox reduced with an excess of sodium dithionite. B (lower spectrum): untreated conditions: microwave frequency, 9418.5 MHz; microwave power incident to the cavity, 2 mW: modulation amplitude, temperature, 35 K. The spectra were recorded with the same receiver gain.

from C. fasciculata proved to contain additional sequence similarity [45], demonstrating that it is the C. fasciculata homologue. No other significant homologies could be found, in line with the fact that the N-terminus of small cox subunits displays very little sequence conservation between different organisms [e.g. 46,471 and the considerable evolutionary distance between kinetoplastids and the other species for which small cox subunit sequences have been determined [48]. The N-terminus of the four largest subunits of the cox complex, including the three that are presumed to be encoded by mt DNA, was blocked. The sequence analysis confirmed the assignment of complexes B and B’ that are present in the 850 mM (NH&SO4 fraction and in the crude extract (Fig. 1) as different (di- and monomeric?) forms of cox, since identical N-terminal sequences were found for proteins at corresponding positions. From this we can also conclude that the purification procedure did not lead to the loss of cox subunits (compare Fig. 1A and Fig. 1B with Fig. 2).

55

in Section cox. EPR 1.27 mT:

4. Discussion We purified cytochrome c oxidase from the insect trypanosomatid C. fasciculata with the aid of a methyl hydrophobic interaction column (methyl-HIC) and identified the resulting preparation as such by: (i) reduced-minus-oxidized spectroscopy showing a Soret band with a maximum at 446 nm and an CIband at 605 nm in the absence of a peak around 560 nm (Fig. 4); (ii) EPR spectra demonstrating quite clearly the typical signals for oxidized heme a and the Cu, centre of intact cytochrome c oxidase (Fig. 5); (iii) cytochrome c oxidation-activity measurements revealing a strong preference for its natural substrate (Table 1). and (iv) 1D and 2D PAGE (Fig. 2, Table 3) showing the electrophoretic behaviour and subunit composition characteristic of eukaryotic cytochrome c oxidases ([17,49], one subunit being the homologue of N. crassa subunit V ( = mammalian subunit IV, [44]). Definitive proof for the identity of complex B’ as cox was obtained when it was eluted in the enzymatically active form after separation in the first dimension

56

D. Speijer et al. 1 Molecular and Biochemical Parasitology 79 (1996) 47-59

on a non-denaturing ‘blue native’ gel, as described for bovine cox [17]. The eluted material displayed a clear cyanide-sensitive cytochrome c oxidation activity. C. fasciculata cox reproducibly eluted from the methyl-HIC in two fractions. The ‘high-affinity’ fraction, eluting at low salt, consisted for 80% of complex B’ and contained only small amounts of ATP synthase, the bc, complex and an unknown protein of approximately 40 kDa (cf. Fig. 1B and Fig. 2). The ‘low-affinity’ fraction, eluting at high salt, also contained all other respiratory-chain complexes. This elution behaviour was not due to overloading the column, since the same profile arose when small amounts of material (1 mg protein) were applied to a 25-ml column. At present, we have no clear explanation for this phenomenon, but we speculate that part of cox polymerized or interacted with other components of the respiratory chain via hydrophobic regions of subunits, and that, once associated with other proteins, these regions no longer interacted with Table 3 N-terminal sequences of subunits of cytochrome from C. fasciculata Subunit 5 6 7

a 9 10

c oxidase

N-terminal sequence FFGKGWDNASLDTIFSSML PHADHRKYKIQREEMPn-P HFSDFNDPRF PRPFGVWAPATTLAEYRAR IPNPFAYSFKWVYSMKKEI FY GGDMHSSDRFKAAWDEIPL HM YMLAFNSKAKARPNFGLR GVGYWH-EVYnKPGQsY LHFPISAPPIEIDYLDNDPL EFAVRTEArKwGF

Lower case lettering indicates that the identification of the amino acid residue in question was ambiguous; -. indicates that no residue could be identified. In the case of subunit 7 two isoforms with a slight difference in migration, but with the same N-terminus were occasionally found. The bold amino acids in subunit 8 are conserved between cox from C. fbsciculata and cox V from N. crassa and cox IV from mouse. Residues conserved between subunit 8 and one of the homologues are italicized. Conservative substitutions: D-E; K,R; 1,L.

the methyl-HIC. Polymerization of cox has been observed in other organisms [50,5 l] and C. fascicdata cox seems to multimerize readily, as judged from the fact that complexes of different mobility (B’ and B) are found in the 2D PAGE analysis. These most likely represent the di- and monomeric forms, respectively (Fig. lB, Fig. 2). In spite of the large differences in cox content, the specific activity of the two methyl-HIC fractions is comparable, suggesting a pronounced loss of activity in the purest fraction. This does not seem to be related to the (partial) loss of an important subunit during the purification procedure from mt vesicles to purified complex, since 2D gel patterns of complex B’ look very similar in the two fractions (compare Figs. 1 and 2 and see below). We have found, however, that purified C. fasciculata cox is extremely labile, losing activity relatively rapidly even when stored at O”C, making it rather difficult to perform reliable activity measurements. In accordance with this, pure C. fasciculata cox has an apparent molar absorbance (Fig. 4) of only about 25% of that of bovine cox in spectral analysis with dithionite as electron donor (compared to 24 mM - ‘. cm - ’ at 605 nm for bovine cytochrome c oxidase [18], assuming 80% purity of the cox preparation and the same molecular weight for the two cox’s). Also the specific activity of C. fasciculata cox in the steadystate kinetics is approximately 25% of that of bovine cox (at least under the ionic conditions used in our assay system, results not shown), if the concentration of cox is calculated from protein concentration and estimated purity. However, if (as in Table 1) the absorbance at 605 nm is used as a measure of cox concentration, which reflects active cox only, the specific activity of C. fasciculata and bovine cox is comparable. The subunit pattern of cox from different preparations was very similar, apart from the difficulties mentioned above with band resolutions for 213 and the occasional appearance of doublet bands [e.g. band 5 (Fig. lB), band 7 (not shown) and band 9 (Fig. 2)]. The doublets most likely arose from limited proteolytic digestion, given that they were observed more frequently at later stages of the purification (compare Figs. 1 and 2). In addition, the proteins present in the band 7

D. Speijer et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

doublet turned out to have the same N-terminal amino acid sequence. The exact number of subunits of C. fasciculata cox cannot be properly assessed without extensive analysis in functional assays, e.g. of the effect of introducing inactivating mutations into the genes encoding candidate subunits. Since an analysis of this type cannot be performed at present, we defined a cox subunit as a protein that remained physically associated with a cox activity-containing complex during gel filtration/electrophoresis and that was present in such a complex in seemingly stoichiometric amounts. Based upon these considerations, we estimate C. fasciculata cox to have 10 subunits. The N-terminus of the four largest subunits of the complex, three of which are presumed to be mt encoded, appeared to be blocked. Preliminary attempts to obtain internal amino acid sequences following conventional enzymatic and chemical cleavage protocols have failed. This is most likely caused by the low amounts of protein available combined with the (inferred) extreme hydrophobicity of the proteins in question. We are in the process of purifying large amounts of C. fasciculata cytochrome c oxidase to generate sufficient quantities of these proteins, so that we can systematically search for the best cleavage and sequencing protocols. Only with the sequences of the mt encoded subunits of respiratory-chain complexes such as cox in hand, can it be proven that edited RNAs are indeed translated by the trypanosomatid mt ribosomes and can the possibility of alternative editing be investigated.

Acknowledgements

The authors thank Janny van den Burg and Louis Hartog for skilled technical assistance, Dr Paul Sloof, Dr Jan Berden and Daniel Blom for stimulating discussions, Leo Nijtmans for expert advice on the 2D gel system, Professor Dr Jozef van Beeumen and Bart Samyn (University of Gent) for initial protein sequencing experiments and W. van Noppen for expert help in the preparation of the manuscript. This research is supported by the Netherlands Foundation for Chemical Research (SON) and the Foundation

57

for Medical and Health Research (MW) which are subsidized by the Netherlands Foundation for Scientific Research (NWO). S.P.J.A. is indebted to SON, for grants, supplied via NWO, which made the purchase of the Bruker ECS 106 EPR spectrometer possible.

References

111Stuart. K. (1993) RNA editing in mitochondria of African trypanosomes. In: RNA Editing, The Alteration of Protein Coding Sequences of RNA (Benne, R., ed.), pp. 25-52. Ellis Horwood, Chichester, UK. 121Hajduk, S.L., Harris, M.E. and Pollard, V.W. (1993) RNA editing in kinetoplastid mitochondria. FASEB J. 7, 54-63. [31Simpson, L., Maslov. D.A. and Blum, B. (1993) RNA editing in Leishmania mitochondria. In: RNA Editing, The Alteration of Protein Coding Sequences of RNA (Benne, R., ed.), pp. 53-85. Ellis Horwood, Chichester. UK. [41Stuart, K. (1993) The RNA editing process in Trypanosoma brucei. Semin. Cell Biol. 4, 251-260. [51Benne. R. (1994) RNA editing in trypanosomes. Eur. J. Biochem. 221, 9-23. WI Blum, B., Bakalara, N. and Simpson, L. (1990) A model for RNA editing in kinetoplast mitochondria: ‘guide’ RNA molecules transcribed from maxicircle DNA provide the edited information. Cell 60, 189-198. 171 Feagin, J.E., Abraham, J.M. and Stuart, K. (1988) Extensive editing of the cytochrome c oxidase III transcript in Trypanosoma brucei. Cell 53, 413-422. PI Corell, R.A., Feagin, J.E., Riley, G.R., Strickland, T.. Guderian, J.A., Myler, P.J. and Stuart, K. (1993) Trypanosoma brucei minicircles encode multiple guide RNAs which can direct editing of extensively overlapping sequences. Nucleic Acids Res. 21, 4313-4320. [9] Shaw, J.M.. Campbell, D. and Simpson, L. (1989) Internal frameshifts within the mitochondrial genes for cytochrome oxidase subunit II and maxicircle unidentified reading frame 3 of Leishmania tarentolae are corrected by RNA editing: evidence for translation of the edited cytochrome oxidase II mRNA. Proc. Natl. Acad. Sci. USA 86, 6220-6224. [IO] Capaldi, R.A. (1990) Structure and function of cytochrome c oxidase. Annu. Rev. Biochem. 59, 569-596. [ll] Capaldi, R.A. (1990) Structure and assembly of cytochrome c oxidase. Arch Biochem. Biophys. 280, 252262. [12] Tsukihara, T., Aoyama, H.. Yamashita, E.. Tomizaki, T. et al. (1995) Structures of metal sites of oxidized bovine heart cytochrome c oxidase at 2.8 A. Science 269, lO691074.

58

D. Speijer et al. 1 Molecular

and Biochemical

M.W., Thomas, J.W. and Gennis, R.B. (1994) u31 Calhoun, The cytochrome oxidase superfamily of redox-driven proton pumps. Trends Biochem. Sci. 19, 3255330. [I41Kleisen, C.M., Borst, P. and Weijers, P.J. (1975) The structure of kinetoplast DNA. I. Properties of the intact multi-circular complex from Crithidia luciliae. Biochim. Biophys. Acta 390, 155 167. L. and Ray, D.S. (1986) Replication of [I51Birkenmeyer, kinetoplast DNA in isolated kinetoplasts from Crithidia fhsciculata. Identification of minicircle DNA replication intermediates. J. Biol. Chem. 261, 2362-2368. Peterson, G.L. (1979) Review of the folin phenol protein quantitation method of Lowry, Rosebrough. Farr and Randall. Anal. Biochem. 100, 201-220. H. and von Jagow, G. (1991) Blue native 1171Schlgger, electrophoresis for isolation of membrane protein complexes in enzymatically active form. Anal. Biochem. 199, 2233231. J.B., Berden. J.A., Herweijer, M.A. and [I81Sloothaak, Kemp, A. (1985) The use of S-azido ATP and 8-azido ADP as photoaffinity labels of the ATP synthetase in submitochondrial particles: evidence for a mechanism of ATP hydrolysis involving two independent catalytic sites? Biochim. Biophys. Acta 809, 27738. P.J.. Hemrika, W. and Bet-den, J.A. (1989) s91 Schoppink, The effect of deletion of the genes encoding the 40 kDa subunit 11 or the 17 kDa subunit VI on the steady state kinetics of yeast ubiquinol:cytochrome < oxido reductase. Biochim. Biophys. Acta 974, 1922201. WI Sloof. P., Arts, G.J.. van den Burg, J., van der Spek, H. and Benne, R. (1994) RNA editing in mitochondria of cultured trypanosomatids: translatable mRNAs for NADH dehydrogenase subunits are missing. J. Bioenerg. Biomembr. 26, 193-203. c oxidase. I. The Pll van Gelder, B.F. (1966) On cytochrome extinction coefficients of cytochrome a and cytochrome a>. Biochim. Biophys. Acta 118, 34-46. E. and Walasek, O.F. (1967) Cytochrome c P21 Margoliash, from vertebrates and invertebrates. In: Methods in Enzymology (Estabrook, R.W. and Pullman, M.E., eds.), Vol. X, pp. 3399348. Academic Press, New York. B., Jarausch, J.. Hartmann, R. and Merle, P. 1231Kadenbach, (1983) Separation of mammalian cytochrome c oxidase into 13 polypeptides by a sodium dodecyl sulfate - gel electrophoresis procedure. Anal. Biochem. 129, 5 17-521. G.T., Chan, S.I.. ~241Hartzell, C.R.. Beinert. H., Babcock, Palmer, G. and Scott, R.A. (1988) Heterogeneity in an isolated membrane protein. Has the ‘authentic cytochrome c oxidase’ been identified? FEBS Lett. 236, l-4. ( ~251Soulimane. T. and Buse, G. (1995) Integral cytochrome oxidase. Preparation and progress towards a three-dimensional crystallization. Eur. J. Biochem. 227, 588-595. [26] Bisson, R. and Schiavo, G. (1986) Two different forms of cytochrome c oxidase can be purified from the slime mold Dictyostelium discoideztm. J. Biol. Chem. 261, 437334376. [27] Power, SD., Lochrie, M.A., Sevarino, K.A., Patterson, T.E. and Poyton, R.O. (1984) The nuclear-coded subunits

Parasitology

79 (1996) 47-59

of yeast cytochrome c oxidase. J. Biol. Chem. 259, 6564 6570. (281Geier, B.M., Schagger, H., Ortwein, C., Link, T.A., Hagen, W.R., Brandt, U. and von Jagow, G. (1995) Kinetic properties and ligand binding of the eleven-subunit cytochrome-c oxidase from Saccharomyces cereuisiae isolated with a novel large-scale purification method. Eur. J. Biochem. 227, 296-302. J., de Vries, B.F., Sloof. ~91 Hensgens, L.A.M., Brakenhoff, P., Tromp, M.C., van Boom, J.H. and Benne, R. (1984) The sequence of the gene for cytochrome c oxidase subunit I, a frameshift containing gene for cytochrome c oxidase subunit II and seven unassigned reading frames in Trypanosoma brucei mitochondrial maxicircle DNA. Nucleic Acids Res. 12, 7327-7344. J.P.J., Sloof. P.. [301 Benne, R.. van den Burg, J., Brakenhoff, van Boom, J.H. and Tromp, M.C. (1986) Major transcript of the frameshifted cox II gene from trypanosome mitochondria contains four nucleotides that are not encoded in the DNA. Cell 46, 819-826. L. t311Shaw, J.M., Feagin, J.E., Stuart, K. and Simpson, (1988) Editing of kinetoplastid mitochondrial mRNAs by uridine addition and deletion generates conserved amino acid sequences and AUG initiation codons. Cell 52. 401411. 1321van der Spek, H., Speijer, D., Arts, G.J., van den Burg, J., van Steeg, H., Sloof, P. and Benne, R. (1990) RNA editing in transcripts of the mitochondrial genes of the insect trypanosome Critbidia fasciculata. EMBO J. 9, 251-262. c re[331Priest, J.W. and Hajduk. S.L. (1992) Cytochrome ductase purified from Crithidia fasciculata contains an atypical cytochrome c,. J. Biol. Chem. 267, 20186-20195. A.B.P., Muijsers, A.O., Demol, H., 1341 van Kuilenburg, Dekker. H.L. and van Beeumen, J.J. (1988) Human heart cytochrome c oxidase subunit VIII. Purification and determination of the complete amino acid sequence. FEBS Lett. 240, 127-132. W.E., Kastrau. D.H.W.. Steffens, G.C.M., ]351 Antholine. Buse. G., Zumft, W.G. and Kroneck, P.M.H. (1992) A comparative EPR investigation of the multicopper proteins nitrous-oxide reductase and cytochrome (’ oxidase. Eur. J. Biochem. 209, 8755881. S.P.J., Falk. K.-E.. Lanne, B. and 1361Aasa, R., Albracht, Vanngard, T. (1976) EPR signals from cytochrome c oxidase. Biochim. Biophys. Acta 422, 260-272. 1371Albracht, S.P.J. (1980) The prosthetic groups in succinate dehydrogenase: number and stoichiometry. Biochim. Biophys. Acta 612, 11-28. S.P.J. and Leeuwerik. F.J. (1979) [381de Vries, S., Albracht, The multiplicity and stoichiometry of the prosthetic groups in QH2:cytochrome c oxidoreductase as studied by EPR. Biochim. Biophys. Acta 546, 316-333. systems in Kineto1391Hill, G.C. (1976) Electron transport plastida. Biochim. Biophys. Acta 456, 1499193. 1401Pettigrew, G.W., Leaver, J.L., Meyer. T.E. and Ryle. A.P. (1975) Purification. properties and amino acid sequence of

D. @ever atypical

cytochrome

c from

et al. / Molecular and Biochemical Parasitology 79 (1996) 47-59

two protozoa,

cilis and Crithidia oncopelti. Biochem.

Euglena gra-

J. 147, 291-302.

[41] Scarpulla, R.C., Agne, K.M. and Wu, R. (1981) Isolation and structure of a rat cytochrome c gene. J. Biol. Chem. 256. 6480-6486. (421 Evans,

M.J. and

matic cytochrome dogenes demarcate Proc.

Natl. Acad.

[43] van Kuilenburg, Meer, N.M. and

Scarpulla,

R.C.

(1988) The human

so-

c gene: Two classes of processed pseua period of rapid molecular evolution. Sci. USA 85, 9625-9629. A.B.P., van Beeumen. J.J., van der Muijsers, A.O. (1992) Subunits VIIa,

b.c of human cytochrome L’ oxidase. Identification of both ‘heart-type’ and ‘liver-type’ isoforms of subunit VIIa 199.

in

human

heart.

[44] Sachs, M.S., Bertrand, handary, U.L. (1989)

Eur.

J.

Biochem.

203,

193-

H.. Metzenberg, R.L. and RajBCytochrome c oxidase subunit V

gene of Neurospora crassa: DNA sequences, chromosoma1 mapping, and evidence that the cya-4 locus specifies the structural gene for subunit V. Mol. Cell. Biol. 9. 566-577. [45] Speijer,

D., Breek.

C.K.D.,

Muijsers.

A.O.,

Groenevelt,

P.X., de Haan. A. and Benne, R. (1996) The sequence of a small subunit of cytochrome c oxidase from Crithidiu $zsciculata which is homologous IV. FEBS Lett. 381. 123-126.

to mammalian

subunit

59

[46] Rizutto, R., Sandona. D., Capaldi, R.A. and Bisson, R. (1991) Characterization of a cDNA encoding subunit VI of cytochrome c oxidase from the slime mold Dictyosteliam discoidearn. Biochim. Biophys. Acta 1089, 386388. [47] Taanman, J.W. and Capaldi, R.A. (1992) Purification of yeast cytochrome c oxidase with a subunit composition resembling the mammalian enzyme. J. Biol. Chem. 267. 22481-22485. [48] Fernandes, A.P.. Nelson, K. and Beverley, S.M. (1993) Evolution of nuclear ribosomal RNAs in kinetoplastid protozoa: perspectives on the age and origins of parasitism. Proc. Natl. Acad. Sci. USA 90, 11608- 11612. [49] Schggger. H., Cramer, W.A. and von Jagow. G. (1994) Analysis of molecular masses and oligomeric states of protein complexes by blue native electrophoresis and isolation of membrane protein complexes by two-dimensional native electrophoresis. Anal. Biochem. 217, 220-230. [50] Hakvoort. T.B.M., Moolenaar, K.. Lankvelt, A.H.M.. Sinjorgo, K.M.C.. Dekker, H.L. and Muijsers, A.O. (1987) Separation, stability and kinetics of monomeric and dimeric bovine heart cytochrome c oxidase. Biochim. Biophys. Acta 894. 347-354. [51] Capaldi, R.A.. Malatesta. F. and Darley-Usmar, V.M. (1983) Structure of cytochrome c oxidase. Biochim. Biophys. Acta 726. I35 - 148.