A carbon monoxide irreducible form of cytochrome c oxidase and other unusual properties of the “monomeric” shark enzyme

Comp. Biochem. Physiol. Vol. l14B, No. 4, pp. 345-352, 1996 Copyright © 1996 Elsevier Science Inc.

ISSN0305-0491/96/$15.00 PII S0305-0491(96)00031-4 ELSEVIER

A Carbon Monoxide Irreducible Form of Cytochrome c Oxidase and Other Unusual Proper,des of the "Monomeric" Shark Enzyme David E. Holm,* Gerald Godette,-~ Celia Bonaventura,~7 Joseph Bonaventura,~ M. David Boatright,~ Linda L. Pearce§ and Jim Peterson* "DEPARTMENTOF CHEMISTRY,THE UNIVERSITYOF ALABAMA,TUSCALOOSA,AL 35487-0336, U.S.A.; )MARINE BIOMEDICALCENTER, DUKEUNIVERSITYMARINELABORATORY,BEAUFORT,NC 28516, U.S.A.; ~CHEMISTRYDEPARTMENT,UNION UNIVERSITY,JACKSON,TN 38305, U.S.A. AND §SCIENCEDEPARTMENT, JUDSONCOLLEGE,MARION,AL 36756, U.S.A. ABSTRACT. Contrary to previous reports, the functional and spectral properties of "monomeric" shark cytochrome c oxidases are not entirely similar to those of the "dimeric" beef enzyme. Most significantly, unlike the behavior of beef oxidase, the fully oxidized shark enzyme is not reducible by carbon monoxide. Also, preparations of the shark enzyme, isolated at pH 7.8-8.0, lead to more than 60% of the sample always being obtained in a resting form, whereas similarly prepared beef oxidase is very often obtained, both by ourselves and others, exclusively in the pulsed form. Although the electronic absorption, magnetic circular dichroism and electron paramagnetic resonance (EPR) spectra of cytochrome c oxidase obtained from several shark species are similar to those of the beef enzyme, there are some significant differences. In particular, the Soret maximum is at 422 nm in the case of the fully oxidized resting shark oxidases at physiological pH and not 418 nm as commonly found for the beef enzyme. Moreover, the resting shark oxidases do not necessarily exhibit a "g = 12" signal in their EPR spectra. The turnover numbers of recent preparations of the shark enzyme are higher than previously reported and, interestingly, do not differ within experimental uncertainty from those documented for several beef isoenzymes assayed under comparable conditions, come BIOCHEMPHYSIOL114B;4:345-352, 1996. KEY WORDS. Cytochrome oxidase, EPR, MCD, proton pump, shark

INTRODUCTION The recent publication of the crystal structures of cytochrome c oxidase from Paracoccus denitrificans (17) and beef heart (27) has, at last, provided a firm structural basis from which to interpret experimental data concerning this fundamentally important enzyme. Before this, some progress toward understanding the three-dimensional structures of prokaryotic terminal oxidases had been made through mutant studies (8,9,16). Reinterpretation of this body of work in light of the new crystal structures will, no doubt, lead to further useful insight. The bacterial oxidases are thought to represent stripped-down models of eukaryotic enzymes, being very similar to the three mitochondrially encoded subunits of the more complicated system, and future investigations of these molecules will surely continue to lead to a much improved understanding of structure-function relaCorrespondence to: J. Peterson, Department of Chemistry, The University of Alabama, Tuscaloosa, AL 35487-0336, U.S.A. Tel. (205) 348-5957; Fax (205) 348-9104. Abbreviations-EPR, electron pararnagnetic resonance; MCD, magnetic circular dichroism. Received 20 July 1995; revised 4 January 1996; accepted 18 January 1996.

tionships in the enzyme. However, the ultimate goal is to understand cytochrome c oxidase in eukaryotes, where, in addition to the oxidation-reduction and proton pumping activities, the species- and tissue-specific function of the 10 nuclear encoded subunits is still unclear (23,29). To date, the majority of work in this area has been carried out with various preparations of the beef enzyme or whole mitochondria from rat liver. To open a new perspective on the problem, we have investigated the properties of cytochrome c oxidases from an alternate source, namely sharks. Shark cytochrome c oxidases display most of the same basic functional characteristics as the mammalian enzyme (5,31). However, unlike the bovine enzyme, the shark oxidases do not form higher molecular weight aggregates than the 13 subunit "monomer" (10,31) and are, therefore, potentially a simpler experimental system for probing the general micellar solution properties of eukaryotic oxidases than the beef heart enzyme. Until now, the variation in aggregation state was the only significant difference in the properties of mammalian and shark oxidases known. In this article, we report an improved preparation of the enzyme from the heart and skeletal muscle of several shark species, characterized by electronic absorption, electron

Holm et al.

346

paramagnetic resonance (EPR) and magnetic circular dichroism (MCD) spectroscopic measurements. The new preparations of shark cytochrome c oxidase have outstanding activity, independent of species. More interestingly though, they are observed to be functionally distinct from the beef heart enzyme in at least two ways not previously documented.

MATERIALS

AND

METHODS

Instrumentation Electronic absorption spectra were recorded on either Perkin-Elmer 25 or Varian DMS 100 spectrophotometers. Concentrations of cytochrome c oxidase were determined as total heme a, using Es87= 24 mM-Icm -1 for the pyridine hemochromagen (22) or A6604 = 12 mM-lcm -1 (dithionite reduced minus oxidized) for the enzyme (see Results). A Dionex D 110 stopped-flow spectrometer was used for the cyanide binding studies. EPR spectra were obtained using a hybrid instrument consisting of a Varian E109E console, used to provide the field modulation to a Bruker B-E 25 magnet, with an ER 082 power supply and B-H 15 field controller, plus a Varian El02 microwave bridge and V-453.3 cylindrical cavity. The spectrometer was fitted with an Oxford Instruments E S R 900 liquid helium flow cryostat. MCD spectra were recorded using an Aviv Associates 41DS circular dichroism spectrometer in conjunction with a Cryomagnetics Incorporated cryomagnet. A "single spectrum" consists of data recorded with the applied field in the forward direction minus the reverse field data, the difference being divided by two. In this manner, contributions arising from natural circular dichroism are subtracted from the spectrum.

Enzyme Preparations METHOD 1. "Yonetani type" cytochrome c oxidase was prepared from the skeletal muscle of both the Atlantic sharpnose shark (Rhizoprionodon terraenovae) and the scalloped hammerhead shark (Sphyrna lewini) by the procedure of Wilson and associates (21,31) at pH 7.4-7.5. KeilinHartree particles were prepared and the membrane solubilized with sodium cholate. Subsequently, the required oxidase was separated from the other membrane proteins by a series of ammonium sulfate fractionation steps and further sodium cholate extraction. The enzyme was finally dispersed in 0.5% (w/w) Tween 80. At no time was any chloride knowingly introduced into the sample. METHOD 2. "Hartzell-Beinert type" samples of the sharpnose shark, blacknose shark (Carcharhinus acronotus) and blacktip reef shark (Carcharhinus limbatus) enzyme were obtained by adopting some of the suggestions of Baker et al. (3) concerning preparations of the bovine enzyme at pH 7 . 8 -

8.0. Importantly, no sodium chloride was added at any stage of these preparations. The modified Hartzell-Beinert procedure was carried out by processing batches of approximately 500 g of frozen red muscle. This was thawed, cut into 1-cm cubes and then put into a blender with 2.0 d 50% (v/v) ice-pH 7.4 buffer (20 mM potassium phosphate, 0.1 mM in EDTA). After high-speed blending for 1 min, the mixture was centrifuged in 500-ml batches (3000 g, 20 min, 5°C). The combined solids were returned to the blender, another 2.0 l icepH 7.4 buffer added and the mixture subjected to additional blending for 1 min. The supernatant was obtained by centrifugation (3000 g, 20 min, 5 °C). The pH of the combined supernatants was checked and adjusted to 7.4, as necessary. At this and all subsequent steps requiring pH adjustment, 3.0 M acetic acid and/or 3.0 M ammonium hydroxide were used. After filtering through cotton wool, the pH of the solution was taken to 5.6. The resulting precipitate (Keilin-Hartree particles) was obtained by centrifugation (3000 g, 20 min, 5°C) and resuspended in 250 ml cold pH 7.4 buffer. After blending for exactly 10 sec, the pH was adjusted to 7.4 and the total heme a content measured spectrophotometrically. Triton X-114, 1.9 g//imol heme a, made up to 60 ml with water, was added to the solution, followed by 45 g ammonium sulfate. Having first ensured all solids were dissolved, the pH was adjusted to 7.4 and the solution allowed to stand at room temperature for 1 hr. After this period, a further 20 g ammonium sulfate was dissolved in the solution, the pH adjusted to 7.8 and the resulting mixture centrifuged (13,000 g, 20 min, 5°C). The precipitate, which was pale green by this stage, was dissolved in 100 ml pH 7.8 buffer (20 mM potassium phosphate, 0.1 mM in EDTA) and the.total heme a content measured spectrophotometrically. Sodium cholate, 0.9 g//lmol heine a, was added and the volume made up to 250 ml with cold water. After the addition of 33 ml saturated {room temperature) ammonium sulfate solution and readjustment of the pH to 7.8, the mixture was allowed to stand at 5°C for 90 min. After this time, the supernatant was obtained by centrifugation (20,000 g, 20 min, 5°C). Further ammonium sulfate fractionation was performed at 45% and 55% saturation (room temperature). The first precipitate thus obtained was discarded; the second, now a brown-green gelatinous substance, contained the oxidase. The precipitate was dissolved in 25 ml pH 7.8 buffer and the total heme a content measured spectrophotometrically. Sodium cholate, 0.9 g//tmol heine a, was added and the volume made up to 50 ml with cold water. After the addition of 25 ml saturated ammonium sulfate solution and readjustment of the pH to 7.8, the mixture was allowed to stand at room temperature for 30 min. The supematant was obtained by centrifugation (20,000 g, 20 min, 5°C), and a further 35 ml saturated ammonium sulfate added to precipitate the purified cytochrome c oxidase, which was collected by centrifugation and redissolved in the following buffer for storage: 50 mM HEPES, 1.0 mM EDTA (Na + salts), pH 7.8 (or pD 8.2), plus 0.5% (w/w) Tween 80 in

Shark Cytochrome c Oxidases

347

TABLE 1. Steady-state activities of cytochrome c oxidase preparations Turnover number ( k .... sec-') Beef**

Kidney Liver Heart Muscle

350 360 390 330 [360 + 30]

Shark$§

380 R. terraenovae 410 R. terraenovae 360 C . acronotus 350 C. limbatus [380 _+ 301

*Sinjorgo et al. (25). J'Potassium phosphate 100 mM, pH 7.4, 1 mM in EDTA, 0.05% lauryl maltoside, 20°C. SPresent work. §Potassium phosphate 100 mM, pH 6.0, 1 mM in EDTA, 0.1% Tween 80, 20°C.

the case of the shark enzyme (only) and 0.1% (w/w) lauryl maltoside for the beef enzyme. The standard spectrophotometric assays of oxidationreduction activity (31) and cyanide binding kinetics (3) were used to characterize preparations. The pulsing protocol consisted of reducing the oxidase with a 50-fold excess of sodium ascorbate in the presence of a catalytic amount of Sigma bovine cytochrome c (10% heine c relative to heme a) and then reoxidizing the enzyme by exposure to air in the presence of a catalytic amount of Sigma bovine catalase (0.1% protoheme relative to heme a). Reaction of oxidized samples with carbon monoxide was carried out anaerobically by maintaining 3-4/zM enzyme under 1 atmosphere CO pressure at 20°C, pH 7.0-7.4. The progress of CO-driven reduction was monitored in a spectrophotometer at 431 nm, the position of the Soret maximum of the partially reduced CO adduct formed. Samples for use in cryogenic near-infrared MCD experiments were prepared in deuterated buffer and diluted 50% (v/v) in d3-glycerol to obtain optical quality glasses upon freezing.

whether pulsed or resting forms of the enzyme were obtained (3,18). In contrast to the behavior of beef heart oxidase at 20°C and pH 7.4-7.8, both types of preparation of the shark enzyme were found to react slowly with added cyanide (Fig. 1 ). Monitoring the reaction at 428 nm, biphasic kinetics were observed and the reaction was typically less than 90% complete after more than 20 min. That is, the shark enzyme was always isolated in a predominantly (60-70%) slow cyanide binding (or "resting") form. After conversion to the pulsed form, the reaction of both types of shark enzyme preparation with cyanide were fast (data not shown), tha~is, greater than 90% complete after less than 2 min. These differences in the behavior of the various enzyme preparations, together with some other functional and spectroscopic characteristics, are summarized in Table 2. Maintained under an atmosphere of carbon monoxide at pH 7.4 and 20°C, Hartzell-Beinert type preparations of the oxidized beef heart enzyme are converted to a partially reduced derivative with a Soret absorption maximum at 431 nm (Fig. 2), whereas, surprisingly, shark cytochrome c oxidase exhibits no evidence of this reaction. Likewise, Yonetani type preparations of the beef enzyme become reduced in the presence of C O (4) by the reaction [Fe(a3)3+-Cul32+] + C O + H 2 0 ---> [Fe(a3)2+-CuB >] + CO2 + 2H +, where the reduced heme a3 then forms a C O adduct. However, once again, the analogous preparations of shark oxi-

I

0.5

I

I

I

I

I

I

Previous studies using Yonetani type (method 1 ) shark cytochrome c oxidase preparations reported turnover numbers in the range of 100-250 sec -1 (moles bovine ferrocytochrome c oxidized/mole enzyme) (10,31). At 20°C (100 mM sodium phosphate, 2 mM sodium EDTA, plus catalase), we find the turnover number of the Hartzell-Beinert (method 2) type shark enzyme to be 380 4- 30 sec -1, with KM = 60 --+ 4)/.zM. Note that under these particular assay conditions of high ionic strength, as recommended by Sinjorgo et al. (25), there is evidently little variation in the turnover number of cytochrome c oxidase prepared from different shark species (Table 1), and furthermore, the measured activity is essentially independent o f p H in the range 6.0-7.5. Cyanide binding kinetics were examined to determine

I

0.0 -0.5

<,
m

m kobs= 0.49 hours I

-1.0

(tl/2 = 1.4 hours)

-1.5

RESULTS

I

-2.0

~ ~ ~ 1 ~'

~

1

-2.5 I

-3.0 0

1

2

3

I

I

4

I

5

Time (minutes)

FIG. 1. Cyanide binding kinetics of oxidized cytochrome c oxidases as isolated. Aerobic conditions, 5/*M e n z y m e con. centration, 0.1 M CN-, 20°C. Beef (0): pH 7.4 in 50 mM

HEPES, 1 mM EDTA (Na salts) 0.1% (w/w) lauryl maltoside. Shark (A): pH 7.4 in 50 mM HEPES, 1 mM EDTA (Na salts) 0.5% (w/w) Tween 80. Reactions monitored at 428 nm (shark) and 433 nm (beef)--close to the maxima in the difference spectra (Cyanide adduct minus e n z y m e as prepared).

348

Holm et al.

TABLE 2. Selected properties of beef and shark cytochrome c oxidases Sorer

Beef Resting (Yonetani)* Pulsed (H.-B.)~ Shark* Resting (Yonetani) Resting (H.-B.) Pulsed (H.-B.)$

EPR g = 12?

C N - binding

CO reducible?

418 424

Yes No

Slow Fast

Yes Yes

422 422 425

Yes No No

Slow Slow Fast

No No No

/~rnax ( n m )

*At pH 7.4 in 50 mM HEPES, 1 mM EDTA (Na salts) 0.5% (w/w) Tween 80. ";'Asprepared (method 2) pH 7.4 in 50raM HEPES, 1 mM EDTA (Na salts) 0.1% (w/w) lauryl maltoside. ~-Prepared from the Hartzell-Beinert (H.-B.) type shark enzymeas described in Materials and Methods).

dase do not undergo this reaction. The CO-driven reduction of the Hartzell-Beinert beef enzyme is clearly biphasic (Fig. 2). The two phases corresponding to initial reduction of the heme a3-CuB pair with a half-time of about 13 min, followed by slower reduction of heme a, in a similar fashion to the previously reported result for the Yonetani enzyme (4). The electronic absorption spectrum of the oxidized shark oxidase remains virtually unchanged when the enzyme is maintained under C O for a period of 2 days at 5oc. If, however, a mild reductant is supplied in the presence of CO, then a C O adduct of the shark enzyme is readily obtained (Fig. 3B). The Soret maximum at 429-431 nm and absence of a large increase in the band intensity around 600 nm indicate this to be an incompletely formed half-reduced derivative, in which heme a and CUA remain oxidized. Figure 3A shows the electronic absorption spectra of the oxidized resting and dithionite reduced forms of the sharpnose shark oxidase. The spectra shown were actually obtained using a Yonetani preparation, but the same data de-

g

A

m 443

=,

\

2s I 400

I 500

/t

7s

(tll2 = 2.8 hours) g

200

700

(nm)

- -

Pulsed

-- -- --

CO-adduct

B

431

I 500

WAVELENGTH

100 150 Time (minutes)

I 600

1

I 400

50

/-~\ 60/.

2

~o 25

0

.

WAVELENGTH

I-0 I--

-2.0 ~ -2.5

Reduced

E

w r

~'o -0.5 • ~" -I .0 ~~-/0 ~'s

Resting

-- -- --

_~- 7s

w

o.o

- -

I 600

! 700

-

-

-1

(nm)

25

FIG. 2. Carbon monoxide-driven reduction of beef cytochrome c oxidase as isolated. Anaerobic conditions, 5 g M enzyme concentration, 1 atmosphere CO, 20°C; pH 7.4 in 50 mM HEPES, 1 mM EDTA (Na salts) 0.1% (w/w) lauryl maltoside. Reaction monitored at 430 n m - - t h e maximum in the partially reduced carbon monoxide adduct.

FIG. 3. Electronic absorption spectra of Atlantic sharpnose shark cytochrome c oxidase derivatives at 20°C, pH 7.4 [50 mM HEPES, 1 mM EDTA (Na salts) 0.5% (w/w) Tween 80], 1 cm pathlength. (A) Fully oxidized resting enzyme ( ) and sodium dithionite excess reduced derivative ( - - ) 8/xM enzyme concentration. (B) Fully oxidized pulsed enzyme ( ) and sodium ascorbate excess reduced CO-bound de. riviative ( - - ) 4 p M enzyme concentration.

Shark Cytochrome c Oxidases

349

|5o( -

-

- - - -

Shork Beef

E

o '7

ca.) <3

soo

o

Ao

~o

~o

,0be

l~

,~o

,do0

,.b0

~o

The EPR spectra of the sharpnose shark enzyme as prepared exhibits differences between the Yonetani (Fig. 5A) and Hartzell-Beinert (Fig. 5B) types. Essentially identical spectra were obtained for the enzyme prepared from other shark sources (data not shown). The strong signal near 1500 G (g = 4.3) in the Yonetani type indicates the presence of a significantly increased amount of adventitious ferric ion compared with that observed in the Hartzell-Beinert type preparations. Importantly, the EPR spectrum of the Hartzell-Beinert type of shark enzyme shows no evidence of a signal below 600 G (g = 12). In this and other respects, inclu4ing sodium chloride dependence, these preparations appeared*to exhibit the same

Wavelength(nm)

FIG. 4. Magnetic circular dichroism spectrum (mean of four scans) of Atlantic sharpnose shark resting cytochrome c oxidase at 2.0 K and 7.0 T, pD 7.8 [25 mM HEPES, 0.5 mM EDTA (Na salts) 0.25% (w/w) Tween 80, 50% (v/v) d3 glycerol], 0.8 m m pathlength, 34 /~M enzyme concentration ( ). MCD spectrum of the beef heart enzyme recorded under the same conditions ( - - - ) .

rived from Hartzell-Beinert type oxidase samples are indistinguishable. The positions of the bands in these spectra are identical to those obtained for the oxidases of other shark species, including the scalloped hammerhead (Sphyrna lewini) enzyme as previously reported (31). However, based on concentration determinations using the pyridine hemochromagen assay, we now find 6.604 25 + 2 mM-Jcm -1 total heme (,/16"604 = 12 -+ 1 mM-lcm -I total heme), that is, rather larger values for these extinction coefficients than those found by iron analysis. We note regarding this point that the previous study (31) did not include any EPR spectroscopy of samples. Consequently, the authors were unaware of the significant amounts of adventitious iron generally found in Yonetani type preparations of the enzyme from sharks (see below). The electronic absorption spectrum of the fully oxidized pulsed and half-reduced C O bound forms of the sharpnose shark oxidase are shown in Fig. 3B for comparison. The near-infrared MCD spectrum of the resting sharpnose shark oxidase at 2.0 K and 7.0 T is presented in Fig. 4 (solid line). Again, the spectrum was actually obtained using a sample prepared by the Yonetani method, but Hartzell-Beinert type samples give spectra that are indistinguishable within experimental uncertainty. The negative feature at 810 nm and the positive feature at 1590 nm are slightly red shifted but otherwise strikingly similar to the same signatures observed in the spectrum of the bovine enzyme (Fig. 4, broken line) and assigned to CuA and heme a, respectively (26). Hendler et al. (15) recently questioned the assignment of the 830-nm absorption feature of the bovine enzyme to Cua. However, at least in the case of MCD transitions, there can be no doubt, whatsoever, that negative features at about 800 nm in the spectra of oxidized cytochrome c oxidase are due almost exclusively to CuA (11).

A

100

I

I

1100

2100

I

3100

4100

5100

Magnetic Field (O)

m_

B

100

I

I

1100

2100

I

3100

4100

5100

Magnetic Field (O) FIG. 5. X-band EPR spectra of fully oxidized Atlantic sharpnose shark cytochrome e oxidase derivatives at pH 7.8 [50 mM HEPES, 1 mM EDTA (Na salts) 0.5% (w/w) Tween 80]. (A) Yonetani preparation at 10 K and 2 mW, 10 G modulation amplitude, 2 x 104 receiver gain, 65 /~M enzyme concentration. (B) Hartzell-Beinert preparation at 12 K and 2 mW, 10 G modulation amplitude, 2 x 104 receiver gain, 80/zM enzyme concentration.

Holm et al.

350

EPR parameters as previously reported for the enzyme from various mammalian and other species (13).

DISCUSSION E n z y m e Activity

Remarkably, at 380 sec -l, the turnover number of the Hartzell-Beinert (method 2) type preparation of shark cytochrome c oxidase (Table 1) does not differ significantly from the maximum values reported for the various bovine isoenzymes (25). This is in keeping with the results of Wilson et al. (30), who found no significant difference between the activities of cytochrome c oxidases isolated from various freshwater (Amazonian) fish and that of the beef enzyme. In addition, Gregory and Ferguson-Miller (12) reported observing essentially the same activity in beef heart and rat liver oxidases reconstituted into phospholipid vesicles. Note that the rate-limiting step in the cytochrome c oxidase catalyzed reduction of molecular oxygen by cytochrome c is the internal electron transfer in the oxidase (2,32) and the measured turnover number is independent of the particular cytochrome c used (31). The measured KM for the shark enzyme is 60/aM, which is actually about three times that reported (25) for the bovine enzyme, but this is not surprising, since bovine cytochrome c was also used in the present assays and may reasonably be expected to have a lower affinity for the shark oxidase. That the turnover numbers of all eukaryotic cytochrome c oxidase isoenzymes are so similar, whether isolated from air-breathing mammals, saltwater cartilaginous fish or bony fish from fresh water of low oxygen content, is of considerable interest. The main conclusion to be drawn from this observation, of course, is that the terminal oxidase constitutes a highly conservative system, evolutionarily speaking. Any major adaptations to changing respiratory needs of a species (either the whole organism or a particular tissue) must, therefore, be met at the level of modifications to the oxygen transport apparatus. Although this suggestion has previously been put forward (30), it was based on less data from fewer species.

Reaction with Cyanide It has long been known that beef heart cytochrome c oxidase of the Yonetani type is isolated in a low activity (resting) form, which becomes activated (pulsed) upon turnover (2,32). However, using alternate purification procedures, some authors have reported the direct isolation of the pulsed enzyme (e.g., reference 6). Also, Baker et al. (3) introduced the terms fast and slow oxidase to describe the widely differing cyanide binding kinetics observed for two forms of the fully oxidized enzyme, obtained in varying ratios in different preparations. These two forms can be preferentially generated simply by controlling the pH during the latter stages ofessentiaUy Hartzell-Beinert type preparations. It has been argued (19)

that "resting" is equivalent to "slow" oxidase and "pulsed" is equivalent to "fast" oxidase. All of our data completely support this point of view. Moreover, in our hands, the modified approach to the Hartzell-Beinert preparation of the beef heart enzyme described by Baker et al. (3) always leads to the isolation of a pulsed derivative--again, in full accord with the findings of others. Interestingly, in all the shark cytochrome c oxidase samples examined thus far, the two phase cyanide binding kinetics typical of nominally resting bo'vine oxidase are evident. Despite the fact that Hartzell-B~inert type preparations of the shark enzyme exhibit no significant g = 12 signal (Fig. 5B), the absence of which is often taken to be indicative of the pulsed enzyme (e.g., reference 20), the exclusively fast cyanide binding kinetics associated with the pulsed system (18) are not observed until the preparation has been taken through an oxidation-reduction cycle (i.e., until it has been pulsed). Therefore, the shark enzyme is always isolated in a predominantly resting form in our current procedures and importantly, the presence or absence of a g = 12 EPR signal cannot be used as an indicator of this fact. Such behavior, summarized in Table 2, demonstrates a clear distinction in the properties of cytochrome c oxidase isolated from beef heart and shark skeletal muscle.

Reaction with Carbon Monoxide Neither type of shark cytochrome c oxidase preparation reacts with CO in the fully oxidized form, whereas the (pulsed) beef enzyme clearly exhibits substantial reduction (ca. 50% heme a3) after less than 1 hr of incubation under CO (Fig. 2). Previously, Yonetani (mostly resting) samples of the beef enzyme have been shown to also undergo CO-driven reduction in a similar period of time (4). Young and Caughey (34) investigated several combinations of conditions and beef enzyme preparations in an effort to inhibit the reaction of oxidized cytochrome c oxidase with CO without success. The lack of reactivity with CO in the case of the shark enzyme cannot be ascribed to the presence of chloride, which does inhibit the reaction if present at 0.5 M concentration in partially aerobic samples (19). This is because oxygen was excluded from the current samples and any residual chloride can only have been present at concentrations many orders of magnitude less than that required to effectively suppress CO-driven reduction. A CO adduct of the shark oxidase is only obtained if some additional reductant, such as ascorbate, is also present (Fig. 3B). Consequently, this represents another sharp contrast in the behavior of oxidases isolated from shark species compared with the beef enzyme. Note that it is seemingly not a question of whether a resting or pulsed derivative is involved, because our Hartzell-Beinert shark samples are 60-70% resting and 3040% pulsed. In the presence of CO, the heme a3 and CuB centers in cytochrome c oxidase have known midpoint potentials of at least +230 mV (1,14). Because the standard reduction

Shark Cytochrome c Oxidases

potential for the reaction COz + 2H + + 2e- --+ CO + HzO is - 104 mV (7), it is untenable to argue that the absence of any observed CO-driven reduction in the case of the shark enzyme could be for thermodynamic reasons; there must be an explanation based on reaction kinetics. Ignoring the possibility of complicated mechanisms and assuming that one heme a3-CuB pair is reduced by one CO molecule, there are at least four broad kinds of scenario that might account for the CO-driven reduction of the shark enzyme being extremely slow: one, the active site is inaccessible to ligands like CO due to the protein conformation; two, CO cannot bind to the reaction center due to the existence of a substitution inert coordination sphere; three, the water molecule required for the reaction is excluded from the active site once CO has bound; and four, the reaction requires the assistance of one or two basic groups to accept the protons produced. Because it is naive to suppose that cytochrome c oxidase is not subject to allosteric effects among its subunits, each of the above factors could be modulated by the aggregation state of the system, which is different in the beef and shark enzymes (10). In view of the fact that ligands such as cyanide bind to the active site of both the beef and shark enzyme, exclusion of CO from the active site of only the shark enzyme does not seem very likely. Also, the existence of a coordination sphere that is inert to CO substitution in the shark enzyme appears equally improbable, especially given the general tendency of most cytochrome c oxidase preparations to exhibit heterogeneity associated with the ligand binding site. Furthermore, in keeping with the known ability of terminal oxidases to accept multiple ligands at the site of interaction with substrate oxygen, the recent crystal structures (17,27) do not lend support to the notion that introduction of a small species like CO could ever prevent a water molecule from entering the active site pocket. Therefore, we are confronted with the absence of one or more basic residues in their unprotonated states as the remaining plausible explanation for the observed COirreducibility of the shark enzyme. This is an intriguing possibility, because proton movement into and out of the active site is required for both the oxygen reduction and proton translocation activities of the enzyme. Further studies with beef heart cytochrome c oxidase are presently underway to determine whether the CO-driven reduction can be used as a probe of proton movement in the enzyme.

Spectral Properties An especially noteworthy difference between the shark and bovine enzymes concerns the wavelength maximum of the Soret band in various fully oxidized derivatives. This is at 422 nm in both resting Yonetani (method 1) and HartzellBeinert (method 2) preparations of shark oxidase at pH 7.4, moving to 424-425 nm after a pulsing protocol (Table 2). In contrast, Yonetani and Hartzell-Beinert preparations of the beef enzyme usually exhibit Soret maxima at 417 and

351

424 nm, respectively (20). Note that the reported difference between the Soret bands of the Yonetani type preparations of the beef and shark enzymes is mirrored by the blue shift of the near-infrared MCD spectrum of beef cytochrome c oxidase compared with the shark enzyme (Fig. 4). In fact, fresh Yonetani preparations of the beef oxidase (resting) initially exhibit a Soret maximum at 421-422 nm, which shifts to 417-418 nm upon standing, especially in the presence of detergents like Tween 80, which are known to promote oligomerization (i.e., "dimerization") of, such samples (10 and references therein) but not the sha~k enzyme (10,31). Consequently, we attribute a Soret bared of 417-418 nm to an aggregated ("dimeric") form of the oxidase and not necessarily a low-activity resting (slow cyanide binding) form, as has been suggested (20). The observation that the aggregation state of cytochrome c oxidase may lead to readily measurable spectral shifts in the position of the Soret band is of some significance. Most current hypotheses concerning the mechanism of proton pumping by this enzyme involve ligand switching processes at one or more of the metal centers (e.g., references 24 and 33). These processes can reasonably be expected to be subject to normal acid-base chemical equilibria when the enzyme is solubilized in detergent micelles. Consequently, a search for pH-dependent spectral shifts in cytochrome c oxidase derivatives represents a viable means of elucidating the site of proton translocation. The bovine enzyme is probably of limited use in such experiments, because its pH-dependent change in aggregation state (10) renders the interpretation of any observed spectral shifts subject to ambiguity. Therefore, the shark enzyme is clearly to be considered the experimental system of choice for this kind of investigation.

Multiple Oxidized Forms It has been well documented that eukaryotic cytochrome c oxidases, may exist in more than one detergent-dependent aggregation state (e.g., references 10 and 28). Also, certain commonly encountered ligand adducts, like the various adventitious peroxy derivatives (18) and the chloride derivative (19), both of which we took some trouble to avoid forming in the present study, can lead to further heterogeneity in sample preparations. Unfortunately, the conclusion is now apparently inescapable that, in addition to the detergent and exogenous ligand dependent variants listed above, there are at least four distinct forms of fully oxidized "native" eukaryotic cytochrome c oxidase that can be isolated and not just two. These are the CO-reducible (e.g., beef) and CO-irreducible (e.g., shark) variants, each of which may be either resting (slow cyanide binding) or pulsed (fast cyanide binding). The assistance of Frank J. Schwartz (University of North Carolina, Institute of Marine Sciences) in obtaining shark tissue is gratefully acknowledged. The stopped-flow measurements of cyanide binding kinetics

352

were performed with the assistance and instrumentation of David W. Kraus (University of Alabama at Birmingham). This work was funded by a Grant-in-Aid from the American Heart Association (Alabama Affiliate) (J.P.), Biomedical Research Support Grant S07RR0715114 (J.P.), with technical and facility support from National Institute of Environmental Health Sciences Center Grant ES0-1908 (C.B., J.B. and G.G.).

Holm et al.

18. 19.

20.

References 1. Anderson, J.L.; Kuwana, T.; Hartzell, C.R. Spectroelectrochemical investigations of stoichiometry and oxidation-reduction potentials of cytochrome c oxidase components in the presence of carbon monoxide: the "invisible" copper. Biochemistry 15:3847-3855;1976. 2. Antonini, E.; Brunori, M.; Colosimo, A.; Greenwood, C.; Wilson, M.T. Oxygen 'pulsed' cytochrome c oxidase: functional properties and catalytic relevance. Proc. Natl. Acad. Sci. USA 74:3128-3132;1977. 3. Baker, G.M.; Noguchi, M.; Palmer, G. The reaction of cytochrome oxidase with cyanide. J. Biol. Chem. 262;596-604; 1987. 4. Bickar, D.; Bonaventura, C.; Bonaventura, J. Carbon monoxide-driven reduction of ferric heme and heme proteins. J. Biol. Chem. 259:10777-10783;1984. 5. Bickar, D.; Lehninger, A.; Brunori, M.; Bonaventura, J.; Bonaventura, C. Functional equivalence of monomeric (shark) and dimeric (bovine) cytochrome c oxidase. J. Inorg. Biochem. 23: 365-372;1985. 6. Brandt, U.; Sch~igger, H.; Von Jagow, G. Purification of cytochrome c oxidase retaining its pulsed form. Eur. J. Biochem. 182:705-711;1989. 7. Bratsch, S.G. Standard electrode potentials and temperature coefficients in water at 298.15 K. J. Phys. Chem. Ref. Data 18:1-21;1989. 8. Brown, S.; Moody, A.J.; Mitchell, R.; Rich, P.R. Binuclear centre structure of terminal protonmotive oxidases. FEBS Lett. 316:216-223;1993. 9. Gennis, R.B. Prokaryotic models for raitochondrial cytochrome c oxidase. Biochem. Soc. Trans. 21:992-998;1993. 10. Georgevich, G.; Darley-Usmar, V.; Malatesta, F.; Capaldi, R.A. Electron transfer in monomeric forms of beef and shark heart cytochrome c oxidase. Biochemistry 22:1317-1322;1983. 11. Greenwood, C.; Thomson, A.J.; Barrett, C.P.; Peterson, J.; George, G.N.; Fee, J.A.; Reichardt, J. Some spectroscopic views of the CuA site in cytochrome c oxidase preparations. Ann. N.Y. Acad. Sci. 550:47-52;1988. 12. Gregory, L.C.; Ferguson-Miller, S. Effect of subunit III removal on control of cytochrome c oxidase activity by pH. Biochem. 27:6307-6314;1988. 13. Hartzell, C.R.; Beinert, H. Components of cytochrome c oxidase detectable by EPR spectroscopy. Biochim. Biophys. Acta 368:318-338;1974. 14. Hendler, R.W.; Westerhoff, H.V. Redox interactions in cytochrome c oxidase: From the "neoclassical" toward "modern" models. Biophys. J. 63:1586-1604;1992. 15. Hendler, R.W.; Harmon, P.A.; Levin, I.W. Near infrared spectral changes of cytochrome aa3 during potentiometric titrations. Biophys. J. 67:2493-2500;1994. 16. Hosler, J.P.; Ferguson-Miller, S.; Calhoun, M.W.; Thomas, J.W.; Hill, J.; Lemieux, L.; Ma, J.; Georgiou, C.; Fetter, J.; Shapleigh, J.; Tecklenburg, M.M.J.; Babcock, G.T.; Gennis, R.B. Insight into the active-site structure and function of sitedirected mutants of bacterial cytochrome aa~ and cytochrome bo. J. Bioenerg. Biomem. 25:121-136;1993. 17. Iwata, S.; Ostermeier, C.; Ludwig, B.; Michel, H. Structure

21. 22. 23. 24.

25.

26.

27.

28.

29.

30. 31.

32. 33.

34.

at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376:660-669;1995. Moody, A.J.; Cooper, C.E.; Rich, P.R. Characterization of 'fast' and 'slow' forms of bovine heart cytochrome c oxidase. Biochim. Biophys. Acta 1059:189-207;1991. Moody, A.J.; Richardson, M.; Spencer, J.P.E.; Brandt, U.; Rich, P.R. "CO2-1igated" cytochrome c oxidase: characterization and comparison with the Cl--ligated enzyme. Biochem. J. 302:821-826;1994. Palmer, G. Current issues in the chemistry of cytochrome c oxidase. J. Bioenerg. Biomem. 25:145-151;1993. Ragan, C.I.; Wilson, M.T.; Darley-Usmar, V.M.; Lowe, P.N. In: Darley-Usmer, V.M.; Rickwood, D; W~son, M.T. (eds). Mitochondria: a practical approach. IRL Pre}s; 1987:101-108. Rawlinson, W.A.; Hale, J.H. Prosthetic groups of the cytochromes present in Corynebacterium diphtherium with especial reference to cytochrome a. Biochem. J. 45:247-255;1949. Rohdich, F.; Kadenbach, B. Tissue specific +egulation of cytochrome c oxidase efficiency by nucleotides. Biochemistry 32: 8499-8503;1993. Rousseau, D.L.; Ching, Y.-C.; Wang, J. Proton translocation in cytochrome c oxidase: redox linkage through proximal ligand exchange on cytochrome a> J. Bioenerg. Biomem. 25: 165-175;1993. Sinjorgo, K.M.C.; Durak, I.; Dekker, H.L.; Edel, C.M.; Hakvoort, T.B.M.; van Gelder, B.F.; Muijsers, A.O. Bovine cytochrome c oxidase, purified from heart, skeletal muscle, liver and kidney, differ in the small subunits but show the same reaction kinetics with cytochrome c. Biochim. Biophys. Acta 893:251-258;1987. Thomson, A.J.; Greenwood, C.; Gadsby, P.M.A.; Peterson, J.; Eglinton, D.G.; Hill, B.C.; Nicholls, P. The structure of the cytochrome a~-CuB site of mammalian cytochrome c oxidase as probed by MCD and EPR spectroscopy. J. Inorg. Biochem. 23:187-197;1985. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.; Shinzawa-Itoh, K.; Nakashima, R.; Yaono, R.; Yoshikawa, S. Structures of the metal sites of oxidized bovine heart cytochrome c oxidase at 2.8 A. Science 269:1069-1074; 1995. VanAcken, T.; Foxall-VanAcken, S.; Castleman, S.; Ferguson-Miller, S. Alkyl glycoside detergents: synthesis and applications to the study of membrane proteins. Methods Enzymol. 125:27-35;1986. Van KiJilenburg, A.B.P.; Van Beeumen, J.J.; Van der Meet, N.M.; Muijsers, A.O. Subunits VIIa, b, c of human cytochrome c oxidase. Identification of both 'heart type' and 'liver type' isoforms of subunit VIIa in human heart. Eur. J. Biochem. 203:193-199;1992. Wilson, M.T.; Bonaventura, J.; Brunori, M. Mitochondrial cytochrome content and cytochrome c oxidase activity of some amazonian fish. Comp. Biochem. Physiol. 62A:245-249;1979. Wilson, M.T.; Lalla-Maharajh, W.; Darley-Usmar, V.; Bonaventura, J.; Bonaventura, C.; Brunori, M. Structural and functional properties of cytochrome c oxidases isolated from sharks. J. Biol. Chem. 255:2722-2728;1980. Wilson, M.T.; Peterson, J.; Antonini, E.; Brunori, M.; Colosimo, A.; Wyman, J. A plausible two-state model for cytochrome c oxidase. Proc. Natl. Acad. Sci. USA 78:7115-7118;1981. Woodruff, W.H. Coordination dynamics of heme-copper oxidases. The ligand shuttle and the control and coupling of electron transfer and proton translocation. J. Bioenerg. Biomem. 25:177-187;1993. Young, L.J.; Caughey, W.S. Autoreduction phenomena of bovine heart cytochrome c oxidase and other metalloproteins. J. Biol. Chem. 262:15019-15025;1987.