enzymatic heme oxygenase activity in soluble extracts of the unicellular red alga, cyanidium...

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS

Vol. 235, No. 2, December, pp. 371-384, 1984

Enzymatic Heme Oxygenase Activity in Soluble Extracts of the

Unicellular Red Alga, Cyanidium caldarium’

SAMUEL I. BEALE’ AND JUAN CORNEJO

Lkivision of Biology and Medicine, Brown University, Providence., Rhode Island 02912

Received June 19, 1984, and in revised form August 7, 1984

Extracts of the phycocyanin-containing unicellular red alga, Cganidium caldarium, catalyzed enzymatic cleavage of the heme macrocycle to form the linear tetrapyrrole bilin structure. This is the key first step in the branch of the tetrapyrrole biosynthetic pathway leading to phycobilin photosynthetic accessory pigments. A mixed-function oxidase mechanism, similar to the biliverdin-forming reaction catalyzed by animal cell-derived microsomal heme oxygenase, was indicated by requirements for Oz and a reduced pyridine nucleotide. To avoid enzymatic conversion of the bilin product to phycocyanobilins and subsequent degradation during incubation, mesoheme IX was substituted for the normal physiological substrate, protoheme IX. Mesobiliverdin IXa was identified as the primary incubation product by comparative reverse-phase high- pressure liquid chromatography and absorption spectrophotometry. The enzymatic nature of the reaction was indicated by the requirement for cell extract, absence of activity in boiled cell extract, high specificity for NADPH as cosubstrate, formation of the physiologically relevant IXCZ bilin isomer, and over 75% inhibition by 1 PM Sn- protoporphyrin, which has been reported to be a competitive inhibitor of animal microsomal heme oxygenase. On the other hand, coupled oxidation of mesoheme, catalyzed by ascorbate plus pyridine or myoglobin, yielded a mixture of ring-opening mesobiliverdin IX isomers, was not inhibited by Sn-protoporphyrin, and could not use NADPH as the reductant. Unlike the animal microsomal heme oxygenase, the algal reaction appeared to be catalyzed by a soluble enzyme that was not sedimentable by centrifugation for 1 h at 200,OOOg. Although NADPH was the preferred reductant, small amounts of activity were obtained with NADH or ascorbate. A portion of the activity was retained after gel filtration of the cell extract to remove low-molecular- weight components. Considerable stimulation of activity, particularly in preparations that had been subjected to gel filtration, was obtained by addition of ascorbate to the incubation mixture containing NADPH. The results indicate that C. caldarium possesses a true heme oxygenase system, with properties somewhat different from that catalyzing heme degradation in animals. Taken together with previous results indicating that biliverdin is a precursor to phycocyanobilin, the results suggest that algal heme oxygenase is a component of the phycobilin biosynthetic pathway. o 1984 Academic Press, Inc

The phycobilins are open-chain tetra- harvesting pigments in blue-green, red, pyrroles that are covalently linked to pro- and cryptomonad algae. A structurally teins and function as photosynthetic light- closely related pigment is phytochrome,

which serves as a photoreversible, photo- 1 Supported by NSF Grant PCM-8213948, USDA morphogenetic pigment in higher plants

Grant Bl-CRCR-l-0679, and NIH BRSG Grant 2- and some algae. The structures of the S0’7-RR07085-17. phycobilins [(l); Fig. l] suggest a biosyn-

2 To whom correspondence should be addressed. thetic similarity to biliverdin, which is

371 0003-9861/84 $3.00 Copyright 0 1984 by Academic Press, Inc.

All rights of reproduction in any form reserved.

372 BEALE AND CORNEJO

IN VITRO HEME OXYGENASE FROM Cyan&urn caldarium 373

formed in animals by oxidative ring-opening of heme (2). Intact cells of the unicellular red alga Cyanidium caldarium were able to take up exogenous protoheme (3) and biliverdin (4), and convert them to the chromophore of the major phycobili- protein, phycocyanin. Recently, cell-free extracts of C. caldarium were shown to catalyze the transformation of biliverdin to phycocyanobilin, which was identical to the pigment derived from phycocyanin by methanolysis (5).

Reported attempts to demonstrate enzymatic conversion of heme to biliverdin by extracts of phycobilin-forming organisms have been unsuccessful (6, 7). One possible reason for the lack of success is that biliverdin is unstable in the cell extracts, and is rapidly converted to phycocyanobilin and other products (5). Me- soheme is a good substrate for animal cell-derived microsomal heme oxygenase (8, 9). Exogenous mesoheme is taken up and converted to mesobiliverdin by C. caldarium cells, but further transformation does not occur, and the mesobiliverdin accumulates within the cells (10). It was also found that mesobiliverdin did not serve as a substrate for in vitro phycocyanobilin formation (5). We now report that soluble extracts of C. caldarium contain heme oxygenase activity that can be assayed by measuring the conversion of mesoheme IX to mesobiliverdin IXa.

MATERIALS AND METHODS

Growth of algal cells. Axenic cultures of C. caldarium were grown in a glucose-based medium at 42°C as previously described (4, 5, 11, 12). Initial stocks were a generous gift from R. F. Troxler (Boston University School of Medicine). For the experiments reported, strain CPD was used. This strain arose in our laboratory as a spontaneous mutant in a wild- type cell culture, and was isolated from a single-cell clone. It forms large amounts of both chlorophyll a and phycocyanin equally well in the light or dark (11). Under our growth conditions, cell number dou- bled in 11 to 12 h, and reached a maximum of 3 X lOa/ml. Cells were grown in 60 ml medium contained in 125-ml DeLong flasks, or in 500 ml medium contained in lOOO-ml Erlenmeyer flasks. Experimen- tal cultures were grown from dilute suspensions, and were harvested during exponential growth at a cell density of approximately 2 X lO*/ml. Each 100

ml of suspension yielded approximately 1.0 g cells. Preparatiolz of cell extracts. Cells were harvested

by centrifugation for 2 min at 1’7508 in an IEC tabletop centrifuge, washed twice with deionized water, and resuspended in 100 mM KPOI, pH 7.4, buffer containing 1 mM EDTA. Cells were broken by passage through a French pressure cell at 5000 psi. Unbroken cells and debris were removed by centrifugation for 2 min at 175Og. The supernatant from the low-speed centrifugation was centrifuged at 10,OOOg for 30 min in a Sorval RC-5B refrigerated centrifuge equipped with a Type HB-4 rotor. This supernatant was used for incubations directly or after high-speed centrifugation for 1 h at 200,OOOg in a Beckman L2-65B centrifuge equipped with a Type 65 rotor. In some instances where indicated, the high-speed supernatant was gel-filtered through a 15-cm-long, 2.5-cm-diameter column of Sephadex G-25 that was preequilibrated with 100 mM KPOI, pH 7.4, 1 mM EDTA. The column was eluted at 4 ml/min with more of the same buffer, and the fraction corresponding to the void volume containing the proteins was collected for incubation.

Incubation conditions. Portions of cell extract were incubated at 42°C for 1 h in a solution comprising 100 mM KPOk, pH 7.4,l mM EDTA, 15 FM mesoheme, and 0.6 mM NADPH. Incubations were carried out in the dark, each in a total volume of 10 ml contained in a 125-ml Erlenmeyer flask, with aeration provided by swirling. Each incubation was with cell extract equivalent to approximately 2 g cells, which contained approximately 60 mg protein. Mesoheme was added from a 1 mM solution dissolved in 0.1 N NaOH. In some cases where indicated, 10 mM ascorbate was added at the beginning of the incubation, and 1.9 mM desferrioxamine was added during the last 2 min of incubation. Variations to this procedure are indicated under Results.

Pigment extraction. The pH of the incubation mixture was adjusted to approximately 4.0 with 180 ~1 100% (w/v) aqueous trichloroacetic acid, and the bilins were extracted with 10 ml methylene chloride:l-butanol (21, v/v). The organic phase was washed with 10 ml water, diluted with an equal volume of methylene chloride:ethanol:water (20:19:1, v/v/v), and applied to a IO-mm-long, &mm-diameter

column of DEAE-Sepharose CL-6B (acetate) as described in (13). The column was washed with 20 ml

methylene chloride:ethanol:water (20~191, v/v/v), and then the pigments were eluted with approximately

1 ml of one of the acetic acid-containing HPLC eluants described below.

In the absence of cell extract, 1 nmol of standard bilin was recovered from incubation medium in 70% yield, 5 nmol in 85% yield, and 10 nmol in above 90% yield.

Synthesis of mesobiliverdin IXa Mesobiliverdin I& was prepared according to the method of


McDonagh and Palma (14). Commercial mesobilirubin IXa derived from biliary bilirubin IXa (10 mg) and 9 mg 2,3-dichloro-5,6-dicyanobenzoquinone were dissolved in 25 ml dimethyl sulfoxide and stirred at room temperature for 5 min. Then ‘75 ml ice-cold water was added, and the pigments were extracted into chloroform:l-butanol(2:1, v/v). The crude product was purified by passage through DEAE-Sephar- ose as described above, and then subjected to pre- parative reverse-phase HPLC. The absorption spectrum in 36% (w/v) aqueous HCl:methanol (1:19, v/ v) had maxima at 359 and 685 nm, exactly as reported for mesobiliverdin dimethyl ester in HCl:methanol [1:20, w/v; Ref. (l)]. The absorption spectrum in chloroform had maxima at 367 and 628 nm, and a blue/red peak-height ratio of 3.35, com- pared to reported peak positions at 369 and 631 nm and a peak-height ratio of 3.46 for mesobiliverdin IXa dimethyl ester (15).

Synthesis of mesobilivwd+n IX isomer mixture. Mixed macrocycle ring-opening isomers of mesobiliverdin IX were prepared by coupled oxidation of mesoheme IX according to the method of O’Carra and Colleran (16). Commercial mesoheme IX derived from natural protoheme IX (3 mg) and 30 mg ascorbic acid were dissolved in 5 ml pyridine plus 10 ml water, and were heated for 30 min at 60°C while air was bubbled into the solution. During this time the solution turned from dark brown to dark green. Next, the solution was cooled and 10 ml 5 N

NaOH was added. After 5 min 36% (w/v) aqueous HCl was added slowly with stirring until the pH was lowered to 3.0. Pigments were extracted into chloroform:l-butanol (2:1, v/v) and washed with water, and a small portion purified by DEAE-Se- pharose chromatography as described above.

HPLC purification of mesobilivwrdin Eluate from the DEAE-Sepharose chromatography was purified by HPLC on an Altex Ultrasphere 5-pm ODS column (1 X 25 cm) at 30°C with a Varian Model 5000 liquid chromatograph. The eluant was ethanokglacial acetic acidwater (87:1:12, v/v/v), and the flow rate was 4 ml/min. Bilins were dissolved in HPLC solvent before injection. Injection volume was 1.0 ml or less. Ab- sorbance was monitored at 360 nm, which is near the short-wavelength absorption maximum for mesobiliverdin in this solvent. Fractions corresponding to various separated components were sometimes collected for later spectrophotometry, as described under Results.

In some cases where the high resolving power of the above system was not needed, increased speed and sensitivity were obtained by using a column of lower diameter (4.6 mm X 25 cm), and changing the eluant composition to ethanol:glacial acetic acidwater (83:4:13, v/v/v) flowing at 1.3 ml/min.

Because of aging of the HPLC columns and minor variations in solvent purity, elution times can vary

from one day to another. For this reason, in all comparative HPLC, identity of elution behavior was confirmed by injecting samples containing mixtures of the two components suspected of being identical, and observing the appearance of a single peak that was not broader than the peaks produced by the separately injected components.

Other procedures. Protein quantitation was by the dye-binding method of Bradford (17), with bovine serum albumin as the standard. Quantitation of mesoheme was by spectrophotometry in pyridine, using the absorption coefficient of 140,4OOM-i reported by Fuhrhop and Smith (18). Quantitation of bilins is described under Results. All analytical spectrophotometry was performed on a Cary 219 instrument (Varian).

Chemicals. Mesobilirubin IXa, mesoheme IX, and Sn-protoporphyrin IX were purchased from Porphy- rin Products, Inc. (Logan, Utah). Desferrioxamine mesylate (brand name, Desferal) was a gift from Ciba-Geigy. All other reagents, solvents, and culture medium components were purchased from Sigma or Fisher.

RESULTS

Quantitatim of bilins. Published molar absorption coefficients are available only for the dimethyl ester of mesobiliverdin IXa in chloroform. The values are 54,700 M-’ at 369 nm and 15,800 M-’ at 631 nm (15). For routine work it is inconvenient to measure the absorption of mesobiliverdin in chloroform due to limited solubility of the unesterified bilin free acid in this solvent. Absorption coefficients were therefore derived for the pigment in 36% (w/v) aqueous HCl:methanol (1:19, v/v) and in the HPLC eluant ethanol:glacial acetic acid:water (87:1:12, v/v/v), by com- paring dilute solutions of HPLC-purified, chemically synthesized mesobiliverdin IXcr in all three solvents (Table I). Quantitation of mesobiliverdin could then be performed by measuring the absorbance after elution from the DEAE-Sepharose or after HPLC purification, using the derived absorption coefficient of 35,100 M-l at 675 nm in HPLC eluant. Using known amounts of injected mesobiliverdin, the HPLC detector response was calibrated under both high- resolution (1 X 25-cm column; 4 ml/min flow rate) and high-speed (4.6 mm X 25 cm column; 1.3 ml/min flow rate) HPLC conditions. With the detector set at 360

IN VITRO HEME OXYGENASE FROM Cyanirlium caldarium 375

TABLE I

SPECTROPHOTOMETRIC CHARACTERIZATION OF MESOBILIVERDIN

Solvent

Blue peak wavelength

(nm)

Blue absorption coefficient

(M-‘1

Red peak wavelength

(nm)

Red absorption coefficient

CM-‘)

Blue/red peak height

ratio

Chloroform [dimethyl ester; Ref. (15)]

Chloroform (this study)

Methanol:HCl [dimethyl ester; Ref. (l)]

Methanol:HCl (this study)

HPLC eluant (this study)

369 54,700 631 15,800 3.46

367 ND” 628 ND 3.35

359 NR 685 NR NR

359 78,600 685 43,500 1.81

361 82,200 675 35,100 2.34

’ Abbreviations: ND, not determined; NR, not reported.

nm, the effective absorbances of mesobili- nm, the effective absorbance was 1.41 X lo7 verdin under the high-resolution and high- mol-’ in the high-resolution HPLC system. speed HPLC conditions were 2.04 and 5.05 Identi$cation of the incubation product. X 10’ molll, respectively. The major bilin formed during incubation

The HPLC detector was also calibrated of cell extracts with mesoheme IX was for response to biliverdin. Standard solu- found to have an HPLC elution time (Fig. tions of HPLC-purified biliverdin were 2) and visible absorption spectrum (Fig. quantitated by measuring the absorbance 3) identical to those of authentic, chemi- in 36% (w/v) HCl:methanol (1:19, v/v), tally synthesized mesobiliverdin IXa. It and using the published absorption values was previously shown that mesobiliverdin of 66,200 and 30,800 M-’ at 377 and 696 IXa! can be readily distinguished from nm, respectively, for the pigment in biliverdin IXa and both the Z- and E- HCl:methanol (1:20, w/v) (19). With the ethylidine isomers of phycocyanobilin by HPLC detector monochromator set at 360 these criteria (5).

0.05

E, 0.04

z

-

0.03 2

4 0.02 : k :: 0.01 4

1

a b

- .I o.oo\

1 I I I I I 10 20 30 0 10 20

Eluiion Time (mid

FIG. 2. (a) Reverse-phase HPLC elution profile of incubation product. (b) Elution profile of mixture of incubation product plus standard mesobiliverdin IXa. The small peaks flanking the major peak may be caused by small amounts of contamination by other mesobiliverdin isomers.


FIG. 3. (a) Absorption spectrum of standard mesobiliverdin IXol in 36% (w/v) HCl:methanol (1:19, v/v). (b) Absorption spectrum of incubation product in same solvent. Absorbance scale is multiplied fivefold for the incubation product. Lines are drawn at the peak wavelengths of 359 and 685 nm.

Determination of the specific macrocycle ring-opening isomer produced during the incubation was achieved by comparison of HPLC elution times. The high-resolution HPLC method employed was able to

achieve partial resolution of all four mesobiliverdin IX isomers formed by coupled oxidation of mesoheme IX in aqueous pyridine (Fig. 4). The elution peak of the IXa isomer coincided exactly with the peak of the incubation product (Fig. 2), and also with the first peak of the mixed isomers (Fig. 4). Therefore, by the combination of spectrophotometry and HPLC, the incubation product was unequivocally identified as the IXa isomer of mesobiliverdin.

Catalytic requirements. With incubations having NADPH as the only electron donor present, maximum activity was obtained in complete medium containing 100 mM

KPOl, pH 7.4, 1.0 mM EDTA, 0.6 mM

NADPH, and 15 PM mesoheme (Table II). When cell extract was either omitted or heated to 100°C for 3 min prior to addition of the other reaction components, only small amounts of product, attributable to chemical attack on mesoheme, were detected. Removal of O2 by bubbling the solution with Og-free Nz during the incubation severely lowered the product yield. The amount of product formed in the absence of added reductant was 31.8% of that with NADPH added; it is probable that this amount of reaction was due to the presence of some endogenous electron

b

I

Elution Time (min)

FIG. 4. (a) Reverse-phase HPLC elution profile of mixed isomers of mesobiliverdin IX produced by coupled oxidation of mesoheme in aqueous pyridine. (b) Elution profile of mixture of standard mesobiliverdin IXcz plus mixed isomers.

IN VITRO HEME OXYGENASE FROM Cyanidium caldarium 377

TABLE II

INCUBATION REQUIREMENTS

Mesobiliverdin Relative recovered activity

Incubation condition (nmol) (% of control)

Complete (- ascorbate, 62 mg protein) 3.37 100.0

Cell extract heated to 1OO’C for 3 min 0.15 4.6

- Cell extract 0.21 6.2 - Mesoheme 0.05 1.5 Oxygen-free

atmosphere 0.25 7.3 - NADPH 1.07 31.8 - NADPH, + 0.85

mM NADH 1.53 45.2 - NADPH, + 10 mM

ascorbate 1.53 45.2 Complete (+ ascorbate,

46 mg protein) 2.41 100.0 Cell extract heated to

1OO’C for 3 min 0.13 5.3 2-min incubation time 0.16 6.7 - Cell extract 1.21 50.2 - Mesoheme 0.04 1.6 - NADPH 0.68 23.3 - Ascorbate 1.51 62.8

Note. Complete incubation medium consisted of 100 mM KPO,, pH 7.4.1 mht EDTA. 0.6 mM NADPH, 15 PM mesobeme, and cell extract containing the indicated amount of protein. Ten-milliliter volumes were incubated for 1 b at 42°C in the dark in open flasks. In the incubation series containing ascorbate, this component was present at 10 mbi, and 1.9 mM desferrioxamine was added during the last two minutes of incubation.

source present in the cell extract. NADPH was the preferred electron donor: Either NADH (0.85 mM) or ascorbate (10 mM) caused only a slight increase in product over that recovered in the absence of any added reductant.

With incubations carried out in the presence of both NADPH and ascorbate, maximum activity was obtained in complete medium containing 100 mM KP04, pH 7.4, 10 mM ascorbate, 1.0 mM EDTA, 0.6 mM NADPH, and 15 pM mesoheme (Table II). In these incubations, 1.9 mM desferrioxamine was added during the last 2 min to chelate ferric iron and force dissociation of the ferric-bilin reaction product, which has been reported to form when ascorbate serves as the reductant (20). Subsequently, it was found that des-

ferrioxamine had little or no effect in the incubations with cell extract (data not shown). About half as much product formed in the absence of cell extract as in complete incubation mixture, indicating significant nonenzymatic attack of ascorbate on the mesoheme. Much less product accumulated in the presence of boiled cell extract plus ascorbate. Although this could be due to protection of the mesoheme from nonenzymatic attack of ascorbate by some component present in the heat- denatured cell extract, another possible explanation is the simultaneous degradation of bilins in incubation mixture containing cell extract (see below).

In all incubations with undenatured cell extract, whether in the presence or absence of ascorbate, mesobiliverdin IXLV was the only isomer that accumulated in significant quantity.

Differences in relative activity in different preparations of cell extract can be partially attributed to differences in com- pleteness of cell breakage, which is indicated by the different amounts of protein obtained per gram of cells broken.

Partial puti~catim. The heme oxygenase activity in the cell extracts remained in the supernatant fraction after 1 h of centrifugation at 200,OOOg (Table III). The small amount of activity in the high- speed pellet can be attributed to residual supernatant present in the pellet, which was not washed to remove traces of supernatant. The recovered activity, per gram extracted cells, was somewhat higher after high-speed centrifugation. A portion of the activity was recovered with the protein fraction after gel filtration through Sephadex G-25 to remove low- molecular-weight components from the high-speed supernatant. Both the total activity in each fraction and the fraction of the total activity recovered after gel filtration were greater when ascorbate was used in addition to NADPH as reductant substrate.

Time course of bilin formation. Increas- ing amounts of product accumulated during the first 60 min of incubation, although not at a constant rate (Fig. 5). After 60 min, the amount of product recovered


TABLE III

PARTIAL PURIFICATION OF ALGAL HEME OXYGENASE

Mesobiliverdin recovered

Cell fraction

Protein nlll01/ recovered nmol/g mz

(mg/g cells) cells protein

-Ascorbate 10,oOOg supernatant

200,oog supernatant 200,000g pellet 200,000g supernatant

after gel filtration

+Ascorbate 10,OOOg supernatant

200,OOOg supernatant 200,OOOg pellet 200,OOOg supernatant

after gel filtration

52 0.30 0.006 32 1.22 0.038 20 0.03 0.001

31

44 1.63 0.037 30 2.52 0.084 14 0.11 0.008

29

0.35

2.00

0.012

0.068

Note. Incubation conditions were as described for Ta- ble II.

decreased as the rate of degradation ex- ceeded that of formation (see below).

Activity dependence on cell extract concentration. For this experiment, it was necessary to discriminate between concentration dependence on the enzyme components in the cell extract and possible effects of low-molecular-weight factors also present in the cell extract. Therefore, the cell extract was gel-filtered through Sephadex G-25 before incubation, even though this procedure lowered the product yield per gram of extracted cells. At zero or low cell extract concentrations very little product accumulated, but at higher cell extract concentrations product yield became proportional to cell extract concentration (Fig. 6).

Inhibition by Sn-protopwphyrin. Sn- protoporphyrin was a potent inhibitor of algal heme oxygenase. Over 75% inhibition was caused by 1 PM Sn-protoporphyrin at mesoheme concentrations up to 15 PM (Table IV). A relatively greater percentage inhibition was achieved at lower mesoheme concentrations, suggesting that Sn- protoporphyrin acts as a competitive in-

hibitor. In this experiment, high-speed supernatant from the cell extract was gel- filtered through Sephadex G-25 to remove low-molecular-weight components, and 10 mM ascorbate was added to the incubation solutions to create incubation conditions comparable to those of the myoglobin experiment (see below).

Cwpled oxidation of myoglobin. To determine whether any differences could be discerned between the reaction catalyzed by the algal extracts and the coupled oxidation of mesoheme catalyzed by myoglobin, a series of incubations was carried out in which myoglobin was substituted for the cell extract. In all of these incubations, the ferric iron chelator desferrioxamine (1.9 mM) was added during the last 2 min to release bilins that might be bound to iron, as has been reported to occur when ascorbate serves as the reductant in the heme oxygenase reaction (20). The results (Table V) show that, in the

FIG. 5. Product recovery versus incubation time. Cells were broken, and portions of the high-speed supernatant equivalent to approximately 2 g cells were incubated as described in the text, with NADPH as the reductant. Product was quantitated from absorbance peak height values of HPLC effluent measured at 360 nm.

IN VITRO HEME OXYGENASE FROM Cyanidium caldurium 379

Extracted Cells (Q)

FIG. 6. Product recovery versus quantity of cell extract present in incubation. Various amounts of high-speed supernatant from broken cell extract were incubated as described in the text. In this experiment, the cell extract was gel-filtered through Sephadex G-25 to remove small molecules before incubation. Product was quantitated from absorbance peak height values of HPLC effluent measured at 360 nm.

presence of ascorbate, myoglobin can catalyze the coupled oxidation of additional mesoheme molecules in addition to the protoheme prosthetic group already present on the protein. This can be seen by the formation of both biliverdin and mesobiliverdin in the complete reaction mixture. A small amount of mesobiliverdin, but no biliverdin, was formed by nonenzymatic attack of ascorbate on mesoheme in the absence of myoglobin. In contrast to the heat sensitivity of the algal extract- catalyzed reaction, the catalytic activity of myoglobin was relatively resistant to heat denaturation; after heating to 100°C for 5 min, the formation of biliverdin and mesobiliverdin was inhibited by less than 50%. Unlike the reaction catalyzed by algal extract, NADPH could not substitute for ascorbate, and the two reductants added together were no more effective than ascorbate alone. Although biliverdin

IXcv was the only isomer of that bilin detected among the products, a significant fraction of the mesobiliverdin product consisted of ring-opening isomers other than IXa (Fig. 7). Sn-protoporphyrin did not inhibit the myoglobin-catalyzed coupled oxidation reaction, even at concentrations in excess of those which produced severe inhibition of the algal extract-catalyzed reaction. At high concentrations of Sn-protoporphyrin, the yield of bilin was, in fact, somewhat greater than in control incubations. The increase could be attributed to a small degree of coupled oxidation catalyzed by Sn-protoporphyrin itself at high concentrations (Table V).

Loss of bilins during incubation. Five- nanomole quantities of mesobiliverdin were recovered from incubation medium lacking cell extract in 85% yield, even after incubation for 1 h at 42°C with NADPH (data not shown). Similar recoveries were obtained from incubation medium containing bovine serum albumin at 6 mg/ml, a concentration similar to that of cell extract protein used in incubations. However, when like quantities of meso-

TABLE IV

INHIBITION OF HEME OXYGENASE ACTIVITY BY &I-PROTOPORPHYRIN

Mesobiliverdin Sn-proto- recovered porphyrin

Mesoheme concen- Percent- concentration tration age of

(PM) (PM) nmol uninhibited

5 0.0 2.55 100.0 5 0.5 0.52 20.2 5 1.0 0.32 12.5

10 0.0 3.14 100.0 10 0.5 0.86 27.5 10 1.0 0.54 17.2

15 0.0 3.34 100.0 15 1.0 0.74 22.2 15 5.0 0.24 7.2

Note. Incubation conditions were as described for those samples containing ascorbate in Table II.


TABLE V

COUPLED OXIDATIONOF MESOHEME CATALYZEDBY MYOGLOBIN

Total Total biliverdin mesobiliverdin Mesobiliverdin recovered recovered IXcv isomer

Reaction components present nmol nmol recovered nmol

Complete (15 FM myoglobin) 19.2 13.6 9.16 - myoglobin <0.23 2.26 1.03 Myoglobin heated to 100°C for 5 min 11.4 7.45 5.01 - mesoheme 17.6 <0.46 <O.ll + 0.6 mM NADPH 17.3 13.5 8.91 - Ascorbate <0.23 <0.46 <O.ll - Ascorbate, + 0.6 mM NADPH 10.23 <0.46 <O.ll

Complete (5 j&M myoglobin) 3.76 5.85 3.22 + 5 pM Sn-protoporphyrin 4.18 6.98 3.61 + 10 PM Sn-protoporphyrin 4.06 7.41 3.70 + 20 PM Sn-protoporphyrin 4.10 8.02 4.01 - Myoglobin, + 5 pM Sn-protoporphyrin <0.23 2.64 0.98 - Myoglobin, + 10 j.tM Sn-protoporphyrin 10.23 3.55 1.12 - Myoglobin, + 20 j.tM Sn-protoporphyrin 10.23 3.92 1.21

Note. Complete incubation medium consisted of 100 mM KPO,, pH 7.4, 10 mM ascorbate, 1 mM EDTA, 15 PM mesoheme, and the indicated concentration of myoglobin. Other conditions were as described for Ta- ble II.

biliverdin were added, in place of mesoheme, to complete incubation mixture containing cell extract, the recovery was only about 60% after 2 min and about 10% after 60 min. High amounts of loss were observed even if the cell extract was denatured by boiling before incubation. When ascorbate was present in the incubation mixture, the observed loss of added mesobiliverdin was also high, even in the absence of cell extract. Unsuccessful attempts were made to increase product yield by adding 0.1 mM HgClz or CdClz to the incubation mixture, either before or after incubation, in order to cleave thiol adducts of the bilins that might have formed during the incubations (21) (data not shown). Addition of the antioxidant BHA3 to the incubation mixture at 0.1 mM also did not prevent loss of bilin. As indicated by the low percentage recovery of mesobiliverdin added at the beginning of the incubation, the heme oxygenase activities reported, as measured by me-

a Abbreviation used: BHA, 2,[3]-tert-butyl-4-hy- droxyanisole.

sobiliverdin accumulation, represent minimum values, and the actual rates of bilin synthesis are probably much greater than the net recoveries would imply.

In order to determine whether the specific accumulation of the IXa mesobiliverdin isomer during incubation of cell extract with mesoheme could be caused by selective degradation of the other isomers rather than by specific synthesis, cell extract was incubated for 1 h with a mixture of the all four mesobiliverdin IX isomers. The remaining bilins were isolated and analyzed by HPLC for isomer composition. All four isomers were present in propor- tions approximately equal to those of the starting material (data not shown).

Rapid degradation of biliverdin prevents the use of the natural heme oxygenase substrate protoheme for assay of activity in these cell extracts. When protoheme was substituted for mesoheme in the assay, biliverdin could not be detected after incubation, even though considerable loss of substrate occurred. Loss of known amounts of added biliverdin after 60 min of incubation was even greater than the

IN VITRO HEME OXYGENASE FROM Cyanidium caldarium 381

b c d

FIG. ‘7. (a) Reverse-phase HPLC elution profile of coupled oxidation products of mesoheme in the absence of catalyst. (b) Coupled oxidation products of mesoheme catalyzed by myoglobin. (c) Coupled oxidation products of mesoheme catalyzed by myoglobin, plus added mixed mesobiliverdin IX isomers. (d) Coupled oxidation products of mesoheme catalyzed by myoglobin, plus added biliverdin IXol.

loss of added mesobiliverdin (data not

shown).

DISCUSSION

Considerable indirect evidence exists to support the hypothesis that, in phycobilin biosynthesis, bilin formation from heme proceeds via a heme oxygenase reaction similar or identical to that responsible for biliverdin formation in animal heme catabolism. This evidence includes the ability of whole cells to incorporate exogenous protoheme into the chromophore of phycocyanin (3), production of equi- molar quantities of bilin and CO by whole cells (22), and the incorporation of oxygen atoms from two separate Oz molecules into the two bilin lactam rings (23), as was previously shown to occur during biliverdin formation in animals (24). However, before now, there have been no reports of in vitro heme oxygenase activity occurring in extracts of any phycobilin- forming organism. Our results show that soluble extracts of C. caldarium can catalyze bilin formation from heme in a reaction that requires a reductant and 02, indicating a mixed-function type of heme oxygenase reaction like that occurring in animal cell preparations.

Modifications to the standard heme oxygenase assay (19) were required to detect and measure the activity in C. caldarium extracts. The differences in the assay requirements help to explain the previous lack of reported success in detecting algal heme oxygenase.

First, although heme oxygenase is a microsomal enzyme in animal cells (2, 8, 25, 26), little or none of the activity in algal extracts was sedimented by centrifugation at 200,OOOg for 1 h, indicating that, unlike the animal-derived enzyme, algal heme oxygenase is a soluble enzyme. A nonmicrosomal location for algal heme oxygenase might be expected, in view of the fact that some phycobilin-containing organisms, the blue-green algae, are pro- caryotic and do not contain microsomes, as was pointed out by Troxler (27). It is also reasonable to propose that, in eu- caryotic phycobilin-containing cells, the pigments are formed within the plastids, which have not been reported to contain microsomes.

Second, in order to permit significant quantities of bilin product to accumulate during the incubations, it was necessary to substitute mesoheme for protoheme, the normal physiological tetrapyrrole substrate. This was required because bil-


iverdin IXa, the product of protoheme IX ring-opening, is subject to rapid transformation to phycobilins (5), which are then further degraded to unidentified products by these cell extracts (J. Cornejo and S. I. Beale, unpublished observations; the cell extracts also are capable of catalyzing the reduction of biliverdin to bilirubin). It was reported earlier that exogenous heme (3) and biliverdin (5) are converted to phycocyanin chromophore by intact C. caldarium cells. On the other hand, exogenous mesoheme was metabolized only as far as mesobiliverdin by whole cells (10). The mesobiliverdin accumulated and was apparently not metabolized further (10). Also, mesobiliverdin did not serve as a substrate in the in vitro preparation capable of catalyzing transformation of biliverdin to phycocyanobilins (5). Meso- heme was reported to be nearly as good a substrate as protoheme for the animal microsomal heme oxygenase (8), and the reaction mechanism for transformation of mesoheme to mesobiliverdin appears to be identical to that for biliverdin formation from protoheme (9). Thus, mesoheme appears to be a more suitable substrate than the natural substrate, protoheme, for in vitro heme oxygenase assay in cell extracts containing biliverdin-me- tabolizing activity.

Although NADPH was capable of serving as the sole reductant for enzymatic reaction, considerably more activity was observed when both NADPH and ascorbate were present in the incubation mixture. This suggests that there may be inefficient coupling of the reducing potential of the NADPH to the enzymatic reaction. In the animal cell-derived microsomal heme oxygenase systems, the reducing potential of NADPH is coupled to heme oxygenase via a cytochrome reductase. Normally, both enzymes are closely associated in the microsomes (20, 28). When the heme oxygenase is solubilized with the aid of detergents, NADPH is no longer capable of serving as a substrate, and an artificial electron donor such as ascorbate must be substituted unless solubilized cytochrome reductase is also added (26,31). Because NADPH can serve as the electron donor in the completely

soluble algal heme oxygenase system, the oxygenase enzyme must be able to accept NADPH directly, or an intervening cytochrome reductase must also be soluble. In the latter case, the higher activity observed when ascorbate is present in addition to NADPH could indicate inefficient coupling of heme oxygenase to cytochrome reductase, caused by dilution of the enzymes in the incubation mixture, com- pared to intact cells. It was observed that full activity required the presence of both reductants, and the activities with either reductant alone were approximately ad- ditive. The relatively greater stimulatory effect of ascorbate occurring in incubations with gel-filtered cell extract suggests that some low-molecular-weight com- pound present in the crude high-speed supernatant participates in coupling the reducing power of NADPH to the heme oxygenase reaction.

Because heme breakdown to bilins is known to occur nonenzymatically in the presence of oxygen and reductants such as ascorbate (29), it was necessary to establish the enzymatic nature of the reaction occurring in C. calclarium extracts. Several lines of evidence provide support for a true enzymatic catalysis. First, product yield was proportional (although not linearly) to the amount of cell extract present and to incubation time. Second, activity due to the cell extract was abol- ished by boiling the extract for 3 min before incubation. Third, the physiologically relevant type IXCV bilin isomer was the only one formed in significant quantity during incubation with cell extract, in contrast to the mixture of all four ring- opening isomers that are formed during chemically coupled oxidation of hemes (29). The specificity was found to be attributable to specific formation of the IXCZ isomer, rather than selective degradation of the other isomers. Fourth, there was a high degree of specificity for NADPH as the reductant cosubstrate, with only relatively small amounts of product formed when NADH or ascorbate was substituted. Fifth, the product of the reaction was mesobiliverdin, rather than biliverdin, indicating that the reaction involved transformation of the administered mesoheme

IN VITRO HEME OXYGENASE FROM Cyanidium calahrium 383

and was not due to degradation of the protoheme prosthetic group of some cel- lular hemoprotein. Sixth, activity was strongly inhibited by Sn-protoporphyrin, which has been shown to be a competitive inhibitor of animal heme oxygenase both in vitro and in vivo (30, 31).

While the evidence listed above supports an enzymatic reaction, it does not identify the nature of the catalyst. Myoglobin and hemoglobin have been reported to catalyze stereospecific formation of type IXa bilins in the presence of ascorbate under physiological conditions (29). To determine whether the algal heme oxygenase is a specific enzyme, or whether, instead, the reaction might be catalyzed nonspecifically by some algal hemoprotein, a comparative study was carried out on coupled oxidation of myoglobin, a well-studied substrate and catalyst for the reaction. In confirmation of previous reports, it was found that, under the reaction conditions employed for the algal heme oxygenase assay, myoglobin yielded biliverdin in the presence of ascorbate. Moreover, considerable heme turnover occurred, as was indicated by the formation of both biliverdin and mesobiliverdin when mesoheme was provided in addition to myoglobin. The coupled oxidation of myoglobin, however, differed from the reaction catalyzed by algal extract in four significant ways. First, whereas the activity in cell extract was sensitive to heat denaturation, the myoglobin-catalyzed reaction was slowed by less than half after heating the myoglobin to the boiling point for 5 min prior to incubation. Second, when NADPH was the only reductant provided, no detectable bilin was formed from myoglobin, although algal extract was capable of catalyzing bilin formation with NADPH as the only added reductant, even after gel filtration. Third, whereas only the IXa! isomer of mesobiliverdin was formed during incubations with algal extract, significant quantities of other ring-opening isomers were formed with myoglobin as the catalyst. Fourth, the coupled oxidation reaction of myoglobin was not inhibited by Sn-protoporphyrin IX, even at much higher concentrations than those which caused profound inhibition of bilin for-

mation in the algal extracts, and which have been reported to inhibit animal microsomal heme oxygenase (30, 31). These differences between the reaction catalyzed by algal extract and the coupled oxidation of myoglobin strongly indicate that the algal heme oxygenase activity is catalyzed by a specific enzyme that is involved in phycobilin formation.

Although conversion of mesobiliverdin to phycocyanobilin was not observed, significant loss occurred during incubation of known amounts of the bilin with cell extract under the conditions used for assay of heme oxygenase activity. Low bilin recovery, relative to the amount of heme catabolized, has previously been noted in microsomal and reconstituted systems from animal cells (32). One possible con- tributing factor in the loss is nonenzymatic attack on mesobiliverdin by thiol- containing components in the cell extract. Biliverdin has been reported to be attacked at the central methine bridge carbon by thiols (21). Attempts to increase product recovery by postincubation administration of sulfhydryl reagents such as HgClz and CdCl, were unsuccessful. Attempts to increase bilin yield by adding BHA to prevent oxidative attack on the bilins by free radicals generated from ascorbate or components in the cell extract were also unsuccessful. Enzymatic degradation might also contribute to the loss of bilin. Various phycobilin-containing algae have been reported to undergo loss of phycobil- iproteins, including the bilin component, under conditions such as nitrogen deli- ciency (33, 34). Also, heme turnover has been shown to occur in various plant and algal cells (35, 36), even though bilin accumulation has not been reported in these cells. Although the specific mechanisms responsible for the loss of mesobiliverdin in the algal cell extracts has not yet been elucidated, work in progress is directed toward minimizing the loss to allow more detailed study of the quantitative and regulatory aspects of phycobilin formation in vitro. In view of the known loss of product, the activities reported here must be considered minimum values. Loss of product is also probably responsible for the observed nonlinear relationship of


product yield to concentration of algal cell extract in the incubation mixture.

While heme oxygenase is primarily a catabolic enzyme in animal cells, it serves a biosynthetic role in phycobilin-containing organisms. Other major tetrapyrrole products in these organisms are hemes and chlorophylls. Biliverdin formation would be the first step in the branched tetrapyrrole pathway unique to phycobilin synthesis, and is therefore a likely step to be under regulatory control. Examples of occasions when specific regulation of phycobilin formation might be required include coordination of formation rates for stoichiometric synthesis of bilin and apoprotein moieties of the phycobilipro- teins, light-dark regulation in those organisms having a light requirement for photosynthetic pigment formation, and differential formation of chlorophylls and phycobilins in response to varying conditions of illumination and nutritional sta- tus. Further study of the regulatory properties of algal heme oxygenase in vitro will be necessary for an understanding of its role in the regulation of phycobilin formation.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23. ACKNOWLEDGMENTS

We thank R. F. Troxler for supplying the original C. caldarium strains; P. R. Sinclair and K. Anderson for advice; CIBA-GEIGY for a gift of desferrioxamine mesylate; and J. D. Weinstein for critically reading the manuscript.

24.

25.

26.

REFERENCES 27.

1. COLE, W. J., CHAPMAN, D. J., AND SIEGELMAN, H. W. (1967) J. Amer. Chem Sot. 89, 3643- 3645. 28.

2. TENHUNEN, R., MARVER, H. S., AND SCHMID, R. (1968) Proc. Nat1 Acad Sci USA 61, ‘748-755.

3. BROWN, S. B., HOLROYD, J. A., TROXLER, R. F., AND OFFNER, G. D. (1980) B&hem J. 191, 137-147.

29.

30.

4. BEALE, S. I., AND CORNEZJO, J. (1984) Arch Biochem Biophys. 227,279-286.

5. BEALE, S. I., AND CORNEIO, J. (1984) Plant Physid 76, in press.

31.

32.

6. BENNEW, A., AND SIEGELMAN, H. W. (1979) in. The Porphyrins (Dolphin, D., ed.), Vol 6, pp. 493-520, Academic Press, New York.

7. O’HEOCHA, C. (1968) Biochem Sot Symp. 28, Sl- 105.

33.

34. THOMAS, J. (1970) Nature (Landon) 228, 181-183. 35. CASTELFRANCO, P. A., AND JONES, 0. T. G. (1975)

Plant Physid 55,485-490. 8. TENHUNEN, R., MARVER, H. S., AND SCHMID, R. 36. WEINSTEIN, J. D., AND BEALE, S. I. (1983) J. Biol

YOSHIDA, T., NOGUCHI, M., AND KIKUCHI, G. (1981) J. B&hem 90,125-131.

BROWN, S. B., AND HOLROYD, J. A. (1984) Biochem J. 217.265-272.

BEALE, S. I., AND CHEN, N. C. (1983) Plant Physid 71,263-268.

WEINSTEIN, J. D., AND BEALE, S. I. (1984) Plant Physiol. 74,146-151.

OMATA, T., AND MURATA, W. (1980) Photo&em Photobiol 31, 183-185.

MCDONAGH, A. F., AND PALMA, L. A. (1980) B&hem J. 189,193-208.

STOLL, M. S., AND GRAY, C. H. (1977) Biochem J. 163, 59-101.

O’CARRA, P., AND COLLERAN, E. (1970) J. Chrc- matogr. 50, 458-468.

BRADFORD, M. M. (1976) And B&hem. 72,248- 254.

FUHRHOP, J.-H., AND SMITH, K. M. (1975) in Porphyrins and Metalloporphyrins (Smith, K. M., ed.), pp. 757-869, Elsevier, Amsterdam/ New York.

MCDONAGH, A. F. (1978) in The Porphyrins (Dol- phin, D., ed.), Vol 6, pp. 293-491, Academic Press, New York.

YOSHIDA, T., AND KIKUCHI, G. (1978) J. Biol. Chem 253, 4230-4236.

MANITTO, P., AND MONTI, D. (1979) Ezperientia 35,1418-1420.

TROXLER, R. F. (1972) Biochemistry 11, 4235- 4242.

TROXLER, R. F., BROWN, A. S., AND BROWN, S. B. (1979) J. Biol. Chem. 254, 3411-3418.

BROWN, S. B., AND KING, R. F. G. J. (1978) B&hem. J. 150, 565-567.

YOSHIDA, T., AND KIKUCHI, G. (1978) J. Biol Chem 253,4224-4229.

YOSHIDA, T., AND KIKUCHI, G. (1979) J. Biol Chem. 254, 4487-4491.

TROXLER, R. F. (1977) in Chemistry and Physi- ology of Bile Pigments (Berk, P. D., and Berlin, N. I., eds.), pp. 431-454, U. S. Government Printing Office, Washington, D. C.

YOSHINAGA, T., SASSA, S., AND KAPPAS, A. (1982) J. Biol. Chem. 257, 7786-7793.

O’CARRA, P., AND COLLERAN, E. (1969) FEBS Z&t. 5, 295-298.

DRUMMOND, G. S., AND KAPPAS, A. (1981) Proc Natl. Acad Sci. USA 78, 6466-6470.

YOSHINAGA, T., SASSA, S., AND KAPPAS, A. (1982) J. Biol. Chem. 257,7778-7785.

NOGUCHI, M., YOSHIDA, T., AND KIKUCHI, G. (1982) J. Biochem. 91,1479-1483.

ALLEN, M. M., AND SMITH, A. J. (1969) Arch. Mikrobiol 69, 114-120.

Chem 258, 6799-6807. (1969) J. Biol Chem 244, 6388-6394.

enzymatic heme oxygenase activity in soluble extracts of the unicellular red alga, cyanidium...

Documents