A aldehyde andCO - PNAS

Proc. Nati. Acad. Sci. USAVol. 89, pp. 5306-5310, June 1992Biochemistry

A cobalt-porphyrin enzyme converts a fatty aldehyde to ahydrocarbon and CO

(decarbonylase/Botryococcus braunU/solubilization)

MICHAEL DENNIS AND P. E. KOLATTUKUDY*The Ohio State Biotechnology Center, 206 Rightmire Hall, The Ohio State University, Columbus, OH 43210

Communicated by Paul K. Stumpf, March 13, 1992 (received for review February 3, 1992)

ABSTRACT The final step in hydrocarbon biosynthesisinvolves loss ofCO from a fatty aldehyde. This decarbonylationis catalyzed by microsomes from Botyrococcus braunu. Amongthe several detergents tested for solubiliing the decarbonylase,octyl fi-glucoside (0.1%) was found to be the most effective andreleased 65% ofthe enzyme activity in soluble form. FPLC ofthesolubilized enzyme preparation with Superose 6 followed byion-exchange FPLC with Mono Q resulted in 200-fold increasein specific activity with 7% recovery. The purified enzymereleased nearly 1 mol of CO for each mol of hydrocarbon.SDS/PAGE of the enzyme preparation showed two proteinbands of equal intensity at 66 and 55 kDa. The absorptionspectrum of the enzyme with bands at 410 un, 425 um, 580 mn,and 620 un suggests the presence of a porphyrin. Electronmicroprobe analysis revealed that the enzyme contained Co.Purification of the decarbonylase from B. braunii grown in"CoC12 showed that 57Co coeluted with the decarbonylase.These results suggest that the enzyme contains Co that might bepart of a Co-porphyrin, although a corrin structure cannot beruled out. Co-protoporphyrin IX itself caused decarbonylationofoctadecanal at 60'C, whereas the metal ion or protoporphyrinalone, or several other metal porphyrins, did not cause decar-bonylation. These results strongly suggest that biosynthesis ofhydrocarbons is effected by a microsomal Co-porphyrin-containing enzyme that catalyzes decarbonylation of aldehydesand, thus, reveal a biological function for Co in plants.

Aliphatic nonisoprenoid hydrocarbons are ubiquitous in liv-ing organisms in both the plant and animal kingdoms (1, 2).Widespread occurrence of hydrocarbons in animals, accu-mulation of hydrocarbons under pathological conditions (3,4), demonstrated biosynthesis of hydrocarbons in mamma-lian nerve tissue (2), and decreased hydrocarbon synthesisassociated with neurological disorders (5, 6) suggest impor-tant biological functions for this class of simple compounds.On the basis of the results obtained with specifically

labeled precursors in higher plant tissues, it was proposedthat n-hydrocarbons are produced by elongation of a fattyacid followed by the loss of the carboxyl carbon (7-9).Subsequent studies with insects (10) and mammals (2) sup-ported this mechanism for alkane biosynthesis. Microsomalpreparations from plant and animal tissues that generatealkanes have been shown to catalyze elongation offatty acids(11, 12). The nature of the reaction that results in the loss ofthe carboxyl carbon remained obscure as the chemical natureof the immediate precursors of hydrocarbon was unknownuntil recently. Aldehydes with one carbon more than thealkanes were found to accumulate when hydrocarbon syn-thesis was inhibited by thiol compounds such as dithioeryth-ritol (DTE) (13). In cell-free preparations that generate hy-drocarbons, the observation that an aldehyde with one car-

bon more than the hydrocarbon was formed suggests that thealdehyde might be the immediate precursor of hydrocarbons(14). However, the aldehyde was not recognized as anintermediate in alkane synthesis because a plausible mech-anism for conversion of an aldehyde to an alkane was notobvious until the discovery that porphyrin-coordinated Rucomplexes catalyze the conversion of aldehydes to corre-sponding alkanes with the loss ofCO (15, 16). This laboratoryhas demonstrated that particulate preparations from youngpea leaves (17), the uropygial gland of the eared grebe (18),and diapausing flesh flies (19) catalyzed decarbonylation ofaldehydes to yield the corresponding alkanes. However, theenzyme that catalyzes this biochemical reaction had not beensolubilized and purified and, thus, the nature of the enzymeand the mechanism of the reaction remained obscure.The green colonial alga Botryococcus braunii, race A,

produces linear odd-numbered C27, C29, and C31 hydrocar-bons that total up to 32% of the alga's dry weight (20-22).Microsomal preparations from this organism catalyze thedecarbonylation of an aldehyde to alkane (23). In this paper,we report the solubilization and purification of the aldehydedecarbonylase and present data that suggest that this enzymecontains a Co-porphyrin and, thus, reveal a possible biolog-ical function for Co in plants.

MATERIALS AND METHODSChemicals. [1-14C]Octadecanoic acid and cis-9-[9,10-

3H]octadecenoic acid were from DuPont/New England Nu-clear. 57CoC12 was obtained from Amersham. Pyridiniumchlorochromate, chloroplatinic acid, and RhCl3-H2O werefrom Aldrich. LiAlH4 was from Alfa (Ward Hill, MA);Scintiverse was from Fisher Scientific; Triton X-100 wasfrom Pierce. Protoporphyrin IX and derivatives were fromPorphyrin Products (Logan, UT). All other reagents werefrom Sigma and J. T. Baker. The rhodium chelate wassynthesized by the procedure of Monson (24). [9,10-3H]Octadecanoic acid was prepared by hydrogenation ofcis-9-[9,10-3H]octadecenoic acid (25); [9,10-3H]octadecanoland [1-14C]octadecanol were synthesized from the corre-sponding labeled fatty acid by LiAIH4 reduction. [1-14C]- and[9,10-3H]Octadecanol were oxidized to the aldehyde by usingpyridinium chlorochromate (26).B. braunii. Strain A from the Austin Culture Collection

(University of Texas) was grown in 20-liter clear carboys(Nalge). Sterile air enriched with 1% CO2 was sparged intothe vessels at a rate of 5 liters/hr through ceramic air stonesfrom a local aquarium store. The alga was grown in modifiedCHU 13 medium at 2x strength (21, 27) under cool-whitefluorescent light at 250C. Cultures were harvested between 9and 12 days by centrifugation.

Abbreviation: DTE, dithioerythritol.*To whom reprint requests should be addressed at: The Ohio StateBiotechnology Center, 206 Rightmire Hall, 1060 Carmack Road,Columbus, OH 43210.

5306

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 12

, 202

1

Proc. Natl. Acad. Sci. USA 89 (1992) 5307

Solubilization of Aldehyde Decarbonylase. B. braunii [25 g(wet weight)] was homogenized in 50 ml of 0.1 M potassiumphosphate (pH 7.0) containing 0.3 M sucrose for 2 min byusing a microhomogenizer. The homogenate was centrifugedfor 1 min at 2940 x g in an Eppendorf microcentrifuge. Thesupernatant was centrifuged for 90 min at 105,000 x g, andthe microsomal pellet was resuspended in 500 p1 of solubi-lization buffer containing 0.1 M potassium phosphate (pH7.0), 2 mM ascorbate, 0.05 mM DTE, and 0.1% of one of thefollowing detergents: Triton X-100, octyl P-glucoside, Noni-det P40, or deoxycholate. After gentle shaking for 30 min at250C, the suspension was recentrifuged at 105,000 x g for 90min. The clear supernatant was carefully removed, and thepellet was resuspended in 250 p1 of solubilization buffercontaining the respective detergent. Both the supernatantand the resuspended pellet were assayed for decarbonylaseactivity. Protein was determined by the method of Bradford(28). All procedures were done at 40C.Enzyme Purification. B. braunii (10-12 days in culture) was

homogenized as indicated above and the enzyme was solu-bilized with 0.1% octyl P-glucoside. Aliquots (200 p1) of thesolubilized decarbonylase were injected onto an FPLC Su-perose 6 HR 10/30 column (Pharmacia), previously equili-brated with 0.05 M potassium phosphate (pH 7.0) containing2 mM ascorbate, 0.05 mM DTE, and 0.1% octyl ,&glucoside.The proteins were eluted with the same buffer at a flow rateof 0.4 ml/min, and 0.5-ml fractions were collected. Fractionscontaining decarbonylase activity were pooled and injectedonto an FPLC Mono Q HR 5/5 column (Pharmacia), whichhad been previously equilibrated with 0.05 M potassiumphosphate (pH 7.0) containing 2 mM ascorbate, 0.05 mMDTE, and 0.1% octyl /-glucoside. The column was washedwith one bed volume of the 0.05 M potassium phosphate (pH7.0) and proteins were eluted with a 10-ml linear gradient of0.0-1.0 M KCl. The column effluent was monitored at 280nm, and 0.5-ml fractions were collected. Column fractionscontaining decarbonylase activity were pooled and concen-trated by Centricon (Amicon) ultrafiltration. This preparationwas used for all further studies on the enzyme.Enzyme Assays. Alkane synthesis was assayed in a manner

similar to that previously described (18). Assays were doneanaerobically in 16 x 100-mm test tubes containing threepolypropylene cups with filter paper strips that were sealedwith serum stoppers. One cup contained RhCl[(CH6)3Pb totrap the enzymatically released CO. Another cup containedmethylbenzethonium hydroxide to absorb any released C02;this trap does not absorb CO. After flushing the reactiontubes for 2 min with N2, 200 p1 of freshly prepared 12.5%(wt/vol) pyrogallol in 20% (wt/vol) KOH was injected intothe third cup to remove any remaining 02. The reactionmixtures were incubated for 5 min at 25°C before injection ofthe enzyme. Each reaction mixture contained 65 pM [9,10-3H, 1-14C]octadecanal, the octyl glucoside-containing solu-bilization buffer, and enzyme in a total volume of 2.0 ml.After incubation at 25°C for 45 min, 200 p1 of 2 M HCl wasadded. After photolysis the CO and CO2 traps were removedand assayed for 14C (17). The lipids extracted from themixture were separated by TLC, and the isolated alkanefraction was assayed for radioactivity as described (17, 18).When needed, aliquots of the aqueous phase were assayedfor radioactivity. The chain length of the alkane product wasdetermined by radio-gas chromatography (23).

Electrophoresis. SDS/PAGE was carried out with a Hoefermodel SE-400 apparatus. Analytical slab gels (10%6) wereformed with a 3% stacking gel, and a discontinuous buffersystem was used (29). Proteins were silver-stained (30).UV/Visible Spectrophotometry. An absorption spectrum of

the decarbonylase preparation was determined on a HewlettPackard model 8451A diode array spectrophotometer againstthe buffer blank.

Electron Microprobe Analysis. Enzyme samples (25 p1containing =4 pg of protein) were dried under vacuum into2.0-cm-long grooves etched into glass plates. The analysiswas carried out on a PDP LB3 French version ofan AmericanDigital instrument containing four spectrometers. The elec-tron beam was set at 40 pm and the detection limits for themetals were between 50 and 70 ppm. Background valuesobtained with the buffer were subtracted from the valuesobtained with the samples containing the decarbonylase.In Vivo "Co Labeling of Decarbonylase. B. braunii was

grown in a 1.8-liter Fernbach flask behind lead shielding inmodified CHU 13 medium containing 0.5 mCi of 57CoC12 (1 Ci= 37 GBq) under cool-white light. Cells (3.2 g) were harvestedafter 10 days by centrifugation and all remaining enzymepreparations and chromatography were done as with theunlabeled decarbonylase. Fractions obtained from FPLC Su-perose 6 HR 10/30 and Mono Q 5/5 columns were assayed for57Co on a Packard 5400 series Crystal II multidetector ycounter. Labelingwas performed on two batches ofB. braunii.Nonenzymatic Decarbonylation with a Co-Porphyrin. De-

carbonylation of 100 pM [9,10-3H; 1-14C]octadecanal byCo-protoporphyrin IX chloride was performed by substitut-ing 0.1 mM of the porphyrin for the decarbonylase in theenzyme assay as described above with the exception oftemperature that was held at 30, 40, 60, or 800C. Reactionswere allowed to proceed for 45 min and the products,[3H]heptadecane and 14CO, were measured as described forthe enzymatic assays. All other metal porphyrins examinedfor decarbonylation activity were at 0.1 mM and the reactiontemperature was 60'C.

RESULTSSolubilization of Aldehyde Decarbonylase. Decarbonylase

activity was found in particulate preparations from bothplants and animals (17-19, 23). To elucidate the nature of theenzyme that catalyzes this reaction and to elucidate themechanism of this reaction, the enzyme has to be solubilizedand purified. A B. braunii microsomal preparation was foundto be suitable for this purpose. Decarbonylase activity in themicrosomes was highest after 10 days of algal growth.Therefore, in all decarbonylase purification experiments, B.braunii was harvested at 10 ± 2 days. To test whether thealdehyde decarbonylase could be solubilized in an enzymat-ically active form, microsomal preparations were incubatedwith 0.1% Triton X-100, 0.1% octyl ff3glucoside, 0.1% Noni-det P40, or 0.1% deoxycholate. The decarbonylase activityfound in the supernatant after centrifugation at 105,000 x gwas considered to be due to the solubilized enzyme. Octyl(-glucoside gave considerably higher solubilized-enzyme ac-tivity than the other detergents; deoxycholate also solubi-lized a considerable portion of the decarbonylase. Octyl,B3glucoside-solubilized enzyme had a specific activity 40-foldhigher than the crude cell-free preparation. This detergentachieved a 65% solubilization of the enzyme. Therefore, allfuture solubilizations of aldehyde decarbonylase were per-formed using 0.1% octyl 3-glucoside.

Purification of Aldehyde Decarbonylase. Upon FPLC Su-perose 6 gel filtration of the solubilized decarbonylase prep-aration from the microsomes, the decarbonylase activity waseluted in fractions at or near the void volume (Fig. 1). Whenthe partially purified decarbonylase from the gel-filtrationstep was subjected to anion-exchange FPLC on a Mono QHR 5/5 column, no aldehyde decarbonylase activity could bedetected in the proteins that were not retained in the column.A linear increase ofthe ionic strength with KCl resulted in theelution of several proteins from the column. Decarbonylaseactivity was located in fractions corresponding to the secondmajor protein peak. The solubilization and purification pro-cedures used resulted in a >200-fold increase in specificactivity of the decarbonylase with a 7% recovery of total

Biochemistry: Dennis and Kolattukudy

Dow

nloa

ded

by g

uest

on

Dec

embe

r 12

, 202

1

5308 Biochemistry: Dennis and Kolattukudy

0 5 10 15 20

.-I

._

co

r-

[7.0

[3.510

Lo

Fraction number

FIG. 1. Purification of aldehyde decarbonylase solubilized from microsomes ofB. braundi and coelution of 57Co with the decarbonylase. (A)Gel filtration. Portions (200 ,ul) ofthe solubilized decarbonylase were injected onto an FPLC Superose 5 HR 10/30 column (Pharmacia) previouslyequilibrated with the solubilization buffer, the proteins were eluted at a flow rate of 0.4 ml/min, and 0.5-ml fractions were collected. (B)Ion-exchange chromatography. Fractions from the gel-filtration step that contain decarbonylase activity were pooled and injected onto an FPLCMono Q HR 5/5 column (Pharmacia) that had been previously equilibrated with the solubilization buffer. After washing the column with onebed volume of the buffer, a linear gradient of 0.0-1.0 M KCI in a total volume of 10 ml of the same buffer was applied, and 0.5-ml fractionswere collected. Radioactivity of [3H]alkane is reported as dpm (x 10-3) and radioactivity of 57Co is reported as cpm (X 10-2).

activity (Table 1). SDS/PAGE of the purified decarbonylasepreparations showed two silver-stained bands of similarintensities at 66 and 55 kDa (Fig. 2).

Stoichiometry of Enzymatic Aldehyde Decarbonylation. Inview of the fact that, with crude microsomal preparationsfrom B. braunii, very little CO could be recovered because offurther metabolism of this product (23), CO evolution fromthe purified enzyme was examined. Both CO and CO2 weretrapped during the decarbonylation of [9,10_3H, 1-14C]octa-decanal with the purified enzyme. The CO released was nearequimolar with the alkane produced; very little CO2 wasobserved; an average of 0.84 mol of CO was recovered foreach mol of heptadecane formed.

UV/Visible Spectra. The absorbance spectrum of the al-dehyde decarbonylase is shown in Fig. 3. Besides the normalprotein absorbance at 260 and 280 nm, the enzyme showedabsorbance peaks at -410 nm, 425 nm, 580 nm, and 620 nm.These peaks are highly indicative of a porphyrin moiety.

Metals. Since metal ion chelators are known to inhibitdecarbonylase activity (17, 18, 23), we examined the purifiedenzyme for the presence of metal ion. Electron microprobeanalyses of different aldehyde decarbonylase preparationsconsistently showed significant amounts of Co with muchsmaller and variable amounts of other metals (Table 2). Onaverage 1536 ppm of Co was observed, which calculated to1.13 x 10-11 mol ofCo per ,ug of protein. To test whether Coreally is the metal contained in the enzyme, B. braunii wasgrown in medium containing 57CoC12 for 10 days. The mi-crosomal preparation obtained from labeled cells was treatedwith octyl f-glucoside to solubilize the enzyme. The solubi-lized protein contained 50% of the 57Co contained in themicrosomal preparation. Upon gel filtration on an FPLCSuperose 6 HR 10/30 column, the elution profile of decar-bonylase activity coincided with that of 57Co (Fig. 1). When

the 57Co-labeled decarbonylase preparation from the gelfiltration was applied to an FPLC ion-exchange Mono Qcolumn, a small amount of label emerged in the wash thatcontained no decarbonylase activity. Upon elution of decar-bonylase activity by a KCI gradient, 57Co was coeluted withthe enzyme activity (Fig. 1).Nonenzymatic Decarbonylation with a Co-Porphyrin. In

view of the report that a Ru-porphyrin can cause nonenzy-matic decarbonylation (16), we tested whether a Co-porphyrin could also decarbonylate octadecanal. Co-protoporphyrin IX chloride decarbonylated [9,10-3H]octade-canal optimally at 60°C but did not exhibit activity at ambienttemperatures (Fig. 4). Neither protoporphyrin IX alone norCo alone caused decarbonylation of octadecanal (Table 3).The metal chelators o-phenanthroline, 8-hydroxyquinoline,and EDTA inhibited decarbonylation by Co-porphyrin. Ofthe other metal porphyrins examined, Fe-protoporphyrin IXchloride showed 8% ofthe decarbonylation activity observedwith Co-protoporphyrin IX chloride, whereas Mn-, Sn-, andZn-protoporphyrin IX chlorides did not generate detectableamounts of alkane from octadecanal at 60°C.

DISCUSSIONThe mechanism of biosynthesis of hydrocarbons remainedobscure for a long time in spite ofthe fact that this simple classof organic compounds are produced by a wide variety of livingorganisms (1). Although hydrocarbon synthesis was shown toinvolve the loss ofthe carboxyl carbon ofchain-elongated fattyacid (9), direct loss ofthe carboxyl carbon ofa fatty acid is notmechanistically feasible without an electron-withdrawinggroup adjacent to the a-carbon. No intermediates with such agroup has been detected. On the other hand experimentalresults suggested that an aldehyde could be the immediateprecursor of a hydrocarbon (13, 14). The discovery of a

Table 1. Purification of decarbonylase from B. brauniiTotal nmol of

Total protein, alkane per Specific activity, Apparent Apparent yield,Step mg min nmol per min per mg purification factor t

Crude 44.2 0.770 0.017 1.00 100Microsomes 0.429 0.097 0.226 13.3 12.6Octyl 3-glucoside 0.122 0.083 0.680 40.0 10.7FPLC Superose 6 0.059 0.063 1.06 62.3 8.1FPLC Mono Q 0.016 0.057 3.53 208 7.3

B. braunii (10 days old; 14.57 g) was used for this purification. The total decarbonylase activity in the microsomal fractionwas much less than that in the crude extract. However, the 105,000 x g supernatant itselfhad very little activity. It is possiblethat the supernatant stimulated the microsomal enzyme.

Proc. Natl. Acad Sci. USA 89 (1992)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 12

, 202

1

Proc. Natl. Acad. Sci. USA 89 (1992) 5309

-200e.. 4&

97.4

z0C4

rz

- 68

-43

FIG. 2. SDS/PAGE of puri-fled aldehyde decarbonylase. Ananalytical slab gel (10%1) with a 3%stacking gel was used with thediscontinuous buffer system of

-18.4 Laemmli (29) and the proteinswere silver-stained by the proce-

-14.3 dure of Morrissey (30). Molecularmasses in kDa are indicated.

chemical decarbonylation by Ru-porphyrin complex at ele-vated temperature provided a possible mechanism for con-version of an aldehyde to a hydrocarbon (15, 16). With thediscovery ofenzymes that catalyze conversion ofan aldehydeto a hydrocarbon and CO, it became clear how the mechanisticdifficulty of decarboxylating a fatty acid is circumvented bydecarbonylating an aldehyde that was already known to beproduced by microsomes by an acyl-CoA reductase (31).The results presented in this paper show that the nonionic

detergent octyl S3-glucoside best solubilized the decarbonyl-ase from the microsomal fraction. Octyl (3-glucoside gaveconsiderably higher activities than the other detergents usedin this study, indicating possible stimulation ofthe enzyme aswell as solubilization. The purification procedure resulted ina 200-fold increase in specific activity when compared to thecrude cell-free preparation, and an apparent recovery of -7%of the original decarbonylase activity. SDS/PAGE revealedtwo protein bands of similar intensities at 66 and 55 kDa.Therefore, we tentatively conclude that the enzyme mayhave an an43 structure. Under the conditions so far used, thepurified enzyme emerged at the void volume of Superose 6and Sepharose 6B gel-filtration columns (data not shown) inthe presence and absence of0.1% octyl ,B-glucoside, probablyindicating that the protein aggregated; therefore, the molec-ular weight of the native enzyme is not known.Upon independent electron microprobe analyses of alde-

hyde decarbonylase preparations, the only metal that wasfound consistently in significant amounts was Co. The pres-ence of Co in the enzyme is consistent with the previousfindings that metal chelators such as EDTA, o-phenanthro-

0.1

.0

ut.0.0

0.0 , , i

300 400 500 600Wavelength, nm

Table 2. Electron microprobe analysis of decarbonylase fromB. braunii

Amount,Metal ppm

Fe 0Co 1536Ni 156Cu 192Mg 0Mn 259Mo 0Zn 351

Metal value reported is that in sample minus buffer controlcontaining 0.05 M potassium phosphate (pH 7.0) and 0.1% octylP-glucoside in triple-distilled deionized water.

line, and 8-hydroxyquinoline severely inhibited aldehydedecarbonylase from the plant and animal systems thus farexamined (17, 18, 23). That the isotope 57Co copurifies withthe decarbonylase activity upon gel-filtration and ion-exchange chromatography ofenzyme preparations from cellsgrown in 57CoCI2 adds further evidence that a Co ion is acomponent of the decarbonylase. With the assumption thataldehyde decarbonylase is a heterodimer consisting of a66-kDa monomer and a 55-kDa monomer, the metal contentof the enzyme was calculated to be 1.37 mol of Co per mol(121 kDa) of protein.

In addition to the normal absorbance in the 260- and280-nm region for a protein, the aldehyde decarbonylaseexhibited absorbance peaks at 410 nm, 425 nm, 580 nm, and620 nm, -suggesting the presence of a porphyrin. Porphyrinsare known to absorb in the 380- to 420-nm region (Soretbands), and in the 500- to 600-nm region (the Q-bands) (32).The absorbance spectrum of a pure Co-protoporphyrin IXchloride shows four absorbance peaks at 350 nm, 410 nm, 535nm, and 570 nm. Considering the fact that porphyrin absor-bance usually shifts to the higher wavelength upon binding tothe apoprotein (33), the absorption spectrum of the decar-bonylase is consistent with the presence of a Co-porphyrinwithin the enzyme. By assuming that 1 mol (121 kDa) ofenzyme contains 1 mol ofporphyrin, an extinction coefficientfor aldehyde decarbonylase at 410 nm was determined to be3.33 x 104 litercm-'-mol-1. This value is in the same range as

40

300s

'0

0$-06

FIG. 3. Absorption spectrum of purified decarbonylase prepara-

tion (solid line). The spectrum was recorded in 0.05 M potassiumphosphate, pH 7.0/2.0 mM ascorbate/0.05 mM DTE/0.1% octyl,B-glucoside on an HP 8451AS diode array spectrophotometer. Buffercontrol showed no absorption (dashed line).

20 -

10'

30 40 50 60 70 80

Temperature, °C

FIG. 4. Decarbonylation of octadecanal by Co-protoporphyrinIX. Decarbonylation was measured with 100 AM [9,10-3H; 1-14C]oc-tadecanal and 0.1 mM Co-protoporphyrin IX, as was done for theenzymatic reaction at indicated temperatures. Reactions were al-lowed to proceed for 45 min and the products, [3H]heptadecane and14CO, were quantitated as described for the enzymatic assays. Allother metal porphyrins examined for decarbonylation activity wereat 0.1 mM and the reaction temperature was 60°C.

Biochemistry: Dennis and Kolattukudy

Dow

nloa

ded

by g

uest

on

Dec

embe

r 12

, 202

1

5310 Biochemistry: Dennis and Kolattukudy

Table 3. Effects of metal chelators on nonenzymaticdecarbonylation with Co-protoporphyrin IX chloride and acomparison of activity with other metal porphyrins at 60"C

Alkane, CO,Addition(s) pmol pmol

Co-protoporphyrin IX chloride 28 20+ EDTA (10.0 mM) 0 0.5+ o-Phenanthroline (1.0 mM) 2.8 6.0+ 8-Hydroxyquinoline (1.0 mM) 0 0

CoCl2 (alone) 0 0.9Protoporphyrin IX (alone) 0 0Fe-protoporphyrin IX chloride 2.1 2.1Mn-protoporphyrin IX chloride 0 0.8Sn-protoporphyrin IX dichloride 0 1.0Zn-protoporphyrin IX 0 1.0Alkane and CO were measured as pmol generated above the

background control reaction with [9,10-3H, 1-14C]octadecanal alone.

many known porphyrin-containing proteins (33). From thisresult and the Co content of the enzyme indicated above, wetentatively concluded that the decarbonylase consists of oneCo-porphyrin per a,8 pair of subunits. Even though thespectrum resembles that of porphyrin more than that ofcorrin, definite identification of the tetrapyrrole in the en-zyme as a porphyrin must await further physical chemicaland chemical characterization that would require quantitiesof enzyme not currently available.The evidence that suggests that the aldehyde decarbonyl-

ase contains a Co-porphyrin prompted us to explore whetherCo-porphyrin has the inherent capability to cause nonenzy-matic decarbonylation. In fact, Co-porphyrin IX chloridecatalyzed the decarbonylation at 600C, whereas the individ-ual constituents ofthe Co-protoporphyrin IX chloride did notcause decarbonylation. Metal ion chelators, known to inhibitthe enzymatic decarbonylation, also inhibited the nonenzy-matic decarbonylation by Co-porphyrin. The other metalporphyrins explored in this study showed no decarbonylationwith the exception of Fe-protoporphyrin IX chloride, whichdecarbonylated octadecanal at 8% of the rate observed withthe Co-protoporphyrin IX chloride. That this Co-porphyrincan catalyze the same reaction as aldehyde decarbonylase athigher temperatures demonstrates the inherent capacity ofthe prosthetic group in the decarbonylase to participate in thereaction catalyzed by the enzyme. Obviously, the apoproteinenhances the basic capability of the prosthetic group to makedecarbonylation possible at the biological temperature. Cohas very similar coordination chemistry to Ru, and rutheniumdiphosphine ligands are known to chemically decarbonylatealdehydes to hydrocarbons at elevated temperatures (15). Itis probable that the decarbonylase enzyme with its Co-porphyrin catalyzes the decarbonylation by using a mecha-nism similar to that suggested for the decarbonylation by therhodium diphosphine ligands at elevated temperatures (34).The purified aldehyde decarbonylase catalyzed the decar-

bonylation of octadecanal to heptadecane and CO with near1:1 stoichiometry. The crude particulate fraction from whichthe decarbonylase was obtained did not show such a stoichi-ometry (23), because this preparation also catalyzed the con-version ofCO to CO2 with subsequent metabolism ofCO2 intoother cellular components. Thus, the alga is able to utilize thenormally toxic CO produced by the decarbonylation. Such adetoxification was not observed with the particulate prepara-tions from the pea leaves (17). Since the decarbonylase ap-peared to be present in association with the cuticle outside thecell, CO generated would probably escape into the atmospherewithout any adverse effects on the plant. Similarly, in theuropygial gland of the grebe, CO probably would escape

through the secretory channels from this holocrine gland (18)and thus may not require a detoxification mechanism.Co has been known for a long time to be an essential

element for plant growth. However, to our knowledge,enzymic reactions involving Co have not been elucidated inplants until now. Considering the omnipresence of hydrocar-bons in plants and their role in waterproofing to preventdessication, it is probable that one important function of Coin plants is its role in decarbonylation. Co-porphyrins mightbe of wider occurrence. There are several reports on Co-porphyrin-containing proteins in bacteria (35-37), but suchproteins have not been previously detected in plants.We thank David Little of the Geological Sciences Department of

The Ohio State University for his technical assistance with theelectron microprobe analysis. This work was supported in part byGrant GM 18278 from the National Institutes of Health.1. Kolattukudy, P. E. (1976) Chemistry andBiochemistry ofNatural Waxes

(Elsevier, Amsterdam).2. Cassagne, C., Darriet, D. & Bourre, J. M. (1977) FEBS Lett. 82, 51-54.3. Kolattukudy, P. E. (1968) Science 159, 498-505.4. Heckers, H., Melcher, F., Dittmar, K. & Kalinowski, H. 0. (1978) J.

Chromatogr. 146, 90-96.5. Dyck, P. J., Thomas, P. K. & Lambert, E. H. (1975) Peripheral Neu-

ropathy (Saunders, Philadelphia), pp. 49-58.6. Bourre, J. M., Cassagne, C., Larrouquere-Regnier, S. & Darriet, D.

(1977) J. Neurochem. 29, 645-648.7. Kolattukudy, P. E. (1966) Biochemistry 5, 2265-2275.8. Kolattukudy, P. E. (1967) Phytochemistry 6, 963-975.9. Kolattukudy, P. E. (1987) in The Biochemistry of Plants, ed. Stumpf,

P. K. (Academic, New York), Vol. 9, pp. 291-314.10. Chu, A. J. & Blomquist, G. J. (1980) Comp. Biochem. Physiol. B 68,

313-317.11. Kolattukudy, P. E. & Buckner, J. S. (1972) Biochem. Biophys. Res.

Commun. 46, 801-807.12. Agrawal, V. P. & Stumpf, P. K. (1985) Arch. Biochem. Biophys. 240,

154-165.13. Buckner, J. S. & Kolattukudy, P. E. (1973) Arch. Biochem. Biophys.

156, 34-45.14. Bognar, A. L., Paliyath, G., Rogers, L. & Kolattukudy, P. E. (1984)

Arch. Biochem. Biophys. 235, 8-17.15. Domazetis, G., Tarpey, B., Dolphin, D. & James, B. R. (1980) J. Chem.

Soc. Chem. Commun., 939-940.16. Domazetis, G., James, B. R., Tarpey, B. & Dolphin, D. (1981) in

Catalytic Activation of Carbon Monoxide, American Chemical SocietySymposium Series, ed. Ford, P. C. (Am. Chem. Soc., New York), pp.243-252.

17. Cheesbrough, T. M. & Kolattukudy, P. E. (1984) Proc. Natl. Acad. Sci.USA 81, 6613-6617.

18. Cheesbrough, T. M. & Kolattukudy, P. E. (1988) J. Biol. Chem. 263,2738-2743.

19. Yoder, J. A., Denlinger, D. L., Dennis, M. W. & Kolattukudy, P. E.(1992) Insect Biochem., in press.

20. Wolf, F. R. (1983) Appl. Biochem. Biotechnol. 8, 249-260.21. Yamaguchi, K., Nakano, H., Murakami, M., Konosu, S., Nakayama, O.,

Kanda, M., Nakamura, A. & Iwamoto, H. (1987) Agric. Biol. Chem. 51,493-498.

22. Metzger, P., Villarreal-Rosales, E., Casadevall, E. & Coute, A. (1989)Phytochemistry 28, 2349-2353.

23. Dennis, M. W. & Kolattukudy, P. E. (1991) Arch. Biochem. Biophys.287, 268-275.

24. Monson, R. S. (1971) Advanced Organic Synthesis: Methods and Tech-niques (Academic, New York), p. 43.

25. Brown, H. C. & Brown, C. A. (1962) J. Am. Chem. Soc. 84, 1495.26. Agrawal, V. P. & Kolattukudy, P. E. (1978) Arch. Biochem. Biophys.

191, 452-465.27. Casadevall, E., Dif, D., Largeau, C., Gudin, C., Chaumont, D. &

Desanti, 0. (1985) Biotechnol. Bioeng. 27, 286-295.28. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.29. Laemmli, U. K. (1970) Nature (London) 227, 680-685.30. Morrissey, J. H. (1981) Anal. Biochem. 117, 307-310.31. Kolattukudy, P. E., Rogers, L. & Larson, J. D. (1981) Methods En-

zymol. 71, 263-275.32. Gouterman, M. (1978) in The Porphyrins, ed. Dolphin, D. (Academic,

New York), Vol. 3, pp. 12-14.33. Grigg, R. (1978) in The Porphyrins, ed. Dolphin, D. (Academic, New

York), Vol. 5, p. 340.34. Doughty, D. H., McGuiggan, M. F., Wang, H. & Pignolet, L. H. (1979)

in Fundamental Research in Homogeneous Catalysis, ed. Tsutui, M.(Plenum, New York), Vol. 3, pp. 909-919.

35. Moura, J. J. G., Moura, I., Bruschi, M., LeGall, J. & Xavier, A. V.(1980) Biochem. Biophys. Res. Commun. 2, 962-970.

36. Hatchikian, E. C. (1981)Biochem. Biophys. Res. Commun. 103,521-530.37. Battersby, A. R. & Sheng, Z.-C. (1982) J. Chem. Soc. Chem. Commun.,

1393-1394.

Proc. Natl. Acad. Sci. USA 89 (1992)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 12

, 202

1