Novel Type of Heme-Dependent Oxygenase Catalyzes Oxidative Cleavage of Rubber (Poly-cis-1,4-Isoprene)

Abstract

An extracellular protein with strong absorption at 406 nm was purified from cell-free culture fluid of latex-grown Xanthomonas sp. strain 35Y. This protein was identical to the gene product of a recently characterized gene cloned from Xanthomonas sp., as revealed by determination of m/z values and sequencing of selected isolated peptides obtained after trypsin fingerprint analysis. The purified protein degraded both natural rubber latex and chemosynthetic poly(cis-1,4-isoprene) in vitro by oxidative cleavage of the double bonds of poly(cis-1,4-isoprene). 12-Oxo-4,8-dimethyltrideca-4,8-diene-1-al (m/z 236) was identified and unequivocally characterized as the major cleavage product, and there was a homologous series of minor metabolites that differed from the major degradation product only in the number of repetitive isoprene units between terminal functions, CHO-CH₂— and —CH₂-COCH₃. An in vitro enzyme assay for oxidative rubber degradation was developed based on high-performance liquid chromatography analysis and spectroscopic detection of product carbonyl functions after derivatization with dinitrophenylhydrazone. Enzymatic cleavage of rubber by the purified protein was strictly dependent on the presence of oxygen; it did not require addition of any soluble cofactors or metal ions and was optimal around pH 7.0 at 40°C. Carbon monoxide and cyanide inhibited the reaction; addition of catalase had no effect, and peroxidase activity could not be detected. The purified protein was specific for natural rubber latex and chemosynthetic poly(cis-1,4-isoprene). Analysis of the amino acid sequence deduced from the cloned gene (roxA [rubber oxygenase]) revealed the presence of two heme-binding motifs (CXXCH) for covalent attachment of heme to the protein. Spectroscopic analysis confirmed the presence of heme, and approximately 2 mol of heme per mol of RoxA was found.

Natural rubber (NR) is a biopolymer that is synthesized by many plants and some fungi. This polymer has been commercially exploited for more than 100 years by cultivating and tapping the rubber tree (Hevea brasiliensis). NR is a polymer of many isoprene units [poly(cis-1,4-isoprene)] that, after cross-linking of the linear polymer chains by sulfur bridges (vulcanization), has superior physical properties. Despite the development of chemosynthetic rubbers, NR is still a necessary raw material for products such as tires, latex gloves, condoms, and seals.

NR does not accumulate in the environment. Many reports on the biodegradability of rubbers were published during the last century (for recent studies see references 2, 9, 11, 12, and 18 and references therein). Even chemically cross-linked (vulcanized) rubbers have been shown to be biodegradable (3, 7, 17). Two biological strategies for microbial NR degradation have been described so far. (i) A large number of bacteria, most of which belong to the actinomycetes, are able to grow and to produce clearing zones on agar media containing NR latex in the form of a milky opaque emulsion as a carbon source (9). So far, Xanthomonas sp. strain 35Y is the only known gram-negative NR-degrading bacterium belonging to this group (18). (ii) The members of the other group of NR-utilizing bacteria do not produce clearing zones on NR latex agar; rather, they are able to solubilize solid pieces of NR and to use the resulting emulsion as a carbon source. Gordonia polyisoprenivorans and Gordonia westfalica belong to this class of bacteria (11, 13).

The basic molecular mechanism by which rubber is degraded is not known. Tsuchi and coworkers were the first researchers to isolate and identify low-molecular-mass oligo(cis-1,4-isoprene) derivatives with aldehyde and keto end groups from rubber-grown cultures of Xanthomonas and Nocardia species (17, 18). It is assumed that degradation of the polymer backbone is initiated by statistical oxidative cleavage of one double bond in the polymer chain. The resulting low-molecular-mass oligo(cis-1,4-isoprene) derivatives then are further degraded, presumably by β-oxidation. Analysis of NR degradation products produced by Streptomyces coelicolor 1A and Streptomyces griseus 1D after 70 days of growth on latex gloves revealed an oligomer pattern similar to that observed for Xanthomonas sp. However, products with different end groups were detected (2, 3). Since all these studies were performed with undefined culture broth, it is not known whether the products identified were formed in one or more enzymatic steps. To our knowledge, no enzyme involved in rubber degradation has been isolated in an active form or described. Recently, a gene of Xanthomonas sp. whose gene product could be involved in rubber degradation was cloned (8), but a particular function could not be assigned to the gene. In this study we succeeded in purifying an extracellular protein with polyisoprene oxygenase activity and in characterizing the cleavage reaction.

MATERIALS AND METHODS

Bacteria, media, and culture conditions.

Xanthomonas sp. was grown in nutrient broth or in a mineral salts medium described by Tsuchii and Takeda (18) with 0.5% glucose or 0.2% purified rubber latex at 30°C. Latex cultures also contained 0.002% Tween 80 and sometimes contained 0.05% yeast extract. Solid media contained 1.5% agar. Latex agar was prepared by the overlay technique; a bottom layer (∼30 ml) of mineral salts agar in a petri disk was overlaid with the same agar supplemented with 0.2% purified latex from H. brasiliensis (percentage of solid rubber) with or without 0.05% yeast extract, which resulted in an opaque overlay. Colonies of Xanthomonas sp. produced translucent clearing zones upon incubation at 30°C within 2 to 4 days, indicating utilization of the latex.

Rubbers.

Rubber latex was prepared from freshly tapped H. brasiliensis. Crude latex contains approximately 35% rubber and 1 to 1.5% proteins. Latex was purified from soluble proteins by repeated (three times) centrifugation and washing with 0.002% Tween 80. The top layer (cream) from each centrifugation step was used for the next centrifugation step, while the bottom fractions were discarded. Latex was heat sterilized and stored at 4°C. Purified latex was a gift from the Rubber Research Institute of Malaysia.

Purification of rubber oxygenase (RoxA).

Rubber oxygenase was purified at 5°C by using a fast-performance liquid chromatography system consisting of an LCC 500 controller, a 500 pump, a UV-1 monitor, a Rec-482 recorder, and an FRAC autosampler (Pharmacia, Uppsala, Sweden). Cell-free supernatant of latex-grown Xanthomonas sp. cells was concentrated by ultrafiltration (30-kDa cutoff) and passed through a Q-Sepharose column (HP HR16/10; Pharmacia) that was preequilibrated with basic buffer (20 mM ethanolamine-HCl [pH 9.5]) at a flow rate of 1 ml min⁻¹. RoxA was eluted from the column with a linear gradient of 0 to 0.15 M NaCl in basic buffer at a concentration of approximately 15 mM. Fractions showing the characteristic absorption spectrum of RoxA were pooled and, after desalting and changing of the buffer by diafiltration (30-kDa cutoff), were applied to a MonoP column (HR 5/5; Pharmacia) that was preequilibrated with 20 mM 1,3-diaminopropane-HCl (pH 11.0) at a flow rate of 0.5 ml min⁻¹. During elution with a linear pH gradient (Pharmalyte HCl [pH 8.5] 1:60; Pharmacia) peaks with the characteristic spectrum of RoxA were observed. These fractions were pooled and passed through a Superdex 200 column (Superdex 200 Prep-grade; Pharmacia) and eluted with 20 mM phosphate buffer (pH 7.0).

Protein determination.

Routinely, protein determinations were performed by the method of Bradford (4). For determination of the heme content, the concentration of purified RoxA was also determined by the bicinchoninic acid assay at 562 nm by using a commercial kit (Perbio Science, Erembodegem, Belgium) and by determining the absorption at 280 nm with a specific molar absorption coefficient of 153,160 M⁻¹cm⁻¹, which was calculated from the amino acid composition as described by Gill and Hippel (6).

Determination of heme content.

The heme content of purified RoxA was determined by the pyridine ferrohemochromogen test (5). Six hundred microliters of purified RoxA (15 μg/ml, as determined by the bicinchoninic acid assay and by the Gill-Hippel assay [6]) was added to an aqueous solution of alkaline pyridine (final concentrations, 7.5 mM NaOH and 25% pyridine; final volume, 800 μl) and reduced by adding a few crystals of sodium dithionite. The heme content was calculated from the absorption at 551 nm (ɛ, 29.1 mM⁻¹ cm⁻¹).

Determination of carbonyl content.

The carbonyl content of rubber degradation products was determined after formation of 2,4-dinitrophenyl hydrazones (in hexane) as described by Katz and Keeney (10) by using a molar absorbance coefficient of 21,500 M⁻¹ cm⁻¹ at 338 nm.

RoxA assay and peroxidase assay.

The following conditions were used for product analysis after RoxA-catalyzed rubber degradation by thin-layer chromatography (TLC), carbonyl content determination, and high-performance liquid chromatography (HPLC). The reaction mixture (total volume, 1 ml) contained 100 μl of purified RoxA (10 to 15 μg/ml), rubber latex (4 μl of a 35% emulsion), and bis-Tris buffer (200 mM; pH 7.0). The reaction was carried out at 40°C for 3 or 4 h in a test tube closed with Parafilm. The mixture was extracted with ethyl acetate or diethyl ether, dried, dissolved in 100 to 200 μl of methanol, and then subjected to TLC, carbonyl content determination, or HPLC analysis. Mixtures without RoxA and with heat-inactivated RoxA (10 min, 95°C) were used as negative controls. One unit of activity corresponded to 1 μmol of generated carbonyl function per min. Peroxidase activity was assayed at 510 nm as described by Mason et al. (14); RoxA was incubated in 100 mM sodium phosphate buffer (pH 7) containing 5 mM 2,4 dichlorophenol, 3.2 mM 4-aminoantipyrene, and 1 mM hydrogen peroxide. Horseradish peroxidase was used as a positive control.

Analysis of rubber degradation products by two-dimensional TLC.

Ethyl acetate extracts dissolved in methanol were spotted onto TLC plates (Kieselgel 60F254; Merck & Co., Inc.), and each plate was developed with benzene-acetone (20:1) in the first dimension and with chloroform-methanol (9:1) in the second dimension. After evaporation of the solvent the plates were developed with anisaldehyde-H₂SO₄ spray reagent.

Analysis of reaction products by HPLC and HPLC-mass spectrometry (MS).

Degradation products were detected at 210 nm by HPLC analysis (Chromeleon Chromatography Data Systems 4.38 equipped with a Dionex UV7Vis detector, a UVD 170S/340S, a Dionex P 580 pump, and a Dionex Gina 50 autosampler; Dionex, Isstein, Germany) with a Grom-Sil 100 RP-8 column (125 by 4 mm; particle size, 5 μm; Grom, Herrenberg, Germany). The mobile phase consisted of 50% (vol/vol) aqueous methanol, and separation of the samples was carried out with a gradient to 100% (vol/vol) methanol. Liquid chromatography mass spectra were obtained in the negative and positive electrospray ionization (ESI) mode with an HP1100 HPLC system (Hewlett-Packard, Waldbronn, Germany) coupled to a VG Platform II quadrupol mass spectrometer (Micromass, Manchester, United Kingdom). Samples were resolved with the column and mobile phase described above.

Spectral analysis.

Proton nuclear magnetic resonance (¹H-NMR) spectra were obtained with an ARX 500 spectrometer (Bruker, Rheinstetten, Germany) at a nominal frequency of 500.15 MHz. Samples were dissolved in CDCl₃. Chemical shifts (δ) were expressed in parts per million relative to tetramethylsilane as an internal standard.

RESULTS

Purification and properties of rubber oxygenase (RoxA).

Xanthomonas sp. produces clearing zones during growth on opaque mineral salts agar with latex as a carbon source. Apparently, Xanthomonas sp. secretes an activity that is able to degrade latex to water-soluble products (Fig. 1A). In order to establish an in vitro assay for latex degradation and to analyze the latex-degrading activity, Xanthomonas sp. was grown on mineral salts medium with purified latex. After 7 to 9 days of incubation the latex was visibly degraded and/or coagulated. The culture fluid was separated from the cells and remaining latex particles by successive centrifugation and filtration (pore size, 0.2 μm). Macromolecular components were concentrated about 20-fold by ultrafiltration (30-kDa cutoff). When aliquots of fresh latex were added to the cell-free concentrated culture fluid and incubated at 30°C, latex was again visibly coagulated or degraded within 2 days; the control, with buffer instead of culture fluid, remained milky. When the same experiment was repeated with either the flowthrough from the 30-kDa filtration step or the concentrated culture fluid that had been heated to 95°C for 10 min before latex was added, no coagulation or degradation was observed. We concluded that the compound responsible for coagulation or degradation of the latex is a heat-sensitive macromolecule, presumably an enzyme. It was noticed that the concentrate had a light reddish color that did not appear to the same extent in glucose-grown cultures. The concentrate had a single absorption maximum around 406 nm (the range from 400 to 600 nm was tested) that could be shifted upon reduction to ∼418 and ∼550 nm (data not shown). These results are characteristic for heme-containing proteins, and we speculated that a heme-containing protein was responsible for latex degradation. Using 2 liters of cell-free culture fluid of NR-grown Xanthomonas sp. as the starting material, we were able to purify a protein (apparent molecular mass after sodium dodecyl sulfate [SDS]-polyacrylamide gel electrophoresis, 65 ± 7 kDa) to apparent electrophoretic homogeneity by diafiltration and subsequent chromatography on Q-Sepharose and MonoP (chromatofocusing). The purified 65-kDa protein showed the same strong absorption at 406 nm as the concentrated culture fluid (Fig. 1B and C); the oxidized protein had absorption bands at 280 nm (γ band), 356 nm, and 406 nm (Soret band) and weaker and broader bands at 532 nm (β band), 560 to 565 nm (α band), and 672 nm. A molar extinction coefficient of 1.8 × 10⁵ M⁻¹ cm⁻¹ was determined for the absorption at 406 nm, which was similar to the coefficients of diheme enzymes (1). The λ₄₀₆/λ₂₈₀ value, which reflected the purity and spectral characteristics of RoxA, was 1.17 in 20 mM phosphate buffer (pH 6.8). When the purified 65-kDa protein was reduced by Na₂S₂O₄, absorption, bands appeared at 418 nm (Soret band), 522 nm (β band), and 549 and 553 nm (both α bands), which corresponded to a heme-pyridine complex. These data are indicative of a hemoprotein belonging to the cytochrome c group (15). Addition of synthetic oligo(1,4-cis-isoprene) to purified RoxA resulted in a shift of the Soret band from 406 to 409 nm, indicating that the substrate binds to the enzyme at the heme site(s). The optical spectra of purified RoxA did not exhibit an absorption band at 695 nm which would be expected for a heme iron-methionine bond.

FIG. 1.

(A) Clearing zone formation for Xanthomonas sp. after 4 days of incubation at 30°C on opaque latex agar. (B) SDS-polyacrylamide gel electrophoresis of the 65-kDa protein (RoxA) at various stages of purification. Lane M, marker proteins; lane S, concentrated cell-free culture fluid; lane IEX, pool after ion-exchange chromatography on Q-Sepharose; lane IEF, pool after isoelectric focusing on MonoP. (C) Spectra of purified RoxA before (dashed line) and after (solid line) reduction by dithionite. The numbers indicate the observed maxima of the spectra.

A trypsin fingerprint analysis of the isolated 65-kDa protein was performed, and the masses of the peptides generated were determined by matrix-assisted laser desorption ionization—time of flight analysis (data not shown). The values obtained for six randomly isolated peptides were in agreement with the values obtained for an in silico trypsin digest of a gene product encoded by a recently cloned gene of Xanthomonas sp. (8). The cloned gene was assumed to be involved in rubber degradation, but the biochemical function of the protein was not known. Determination of the amino acid sequences of these six peptides confirmed that the 65-kDa protein was identical to the gene product mentioned above (peptide 1, NH₂-YGLYPAPFR; peptide 2, NH₂-TTPITALGNLLPLPWSTGR; peptide 3, NH₂-GLEDEFEDINNFLISLSPATYPK; peptide 4, NH₂-GVAAVVTPIETIR; peptide 5, NH₂-AWNSGWWAYNNLSPSWTGYPSDNIVASELR; peptide 6, NH₂-WALIEYIK). The deduced amino acid sequence of the cloned gene contained two heme-binding motifs (CXXCH) typical for covalently bound heme. Calculation of the heme content of the purified 65-kDa protein by using the molar absorption coefficient of heme-cytochromes (ɛ, 29.1 mM⁻¹ cm⁻¹ at 551 nm [5]) and three independent methods for protein determination revealed heme and protein concentrations of 16.7 and 9 μM, respectively. These values correspond to a heme content of 1.9 mol of heme per mol of the 65-kDa protein and are in good agreement with the gene sequence analysis that predicted two heme-binding sites. An apparent molecular mass of 54 kDa for the purified native protein was determined by gel filtration (Sephadex G-200). Apparently, the protein has a monomeric subunit structure. In most experiments, the SDS-denatured (reduced) protein migrated at values corresponding to an apparent molecular mass of 55 to 65 kDa, which were significantly lower than the theoretical molecular mass of the mature protein (72.9 kDa) deduced from the gene sequence. The reason for the discrepancy in apparent molecular masses is not known. To test whether the purified protein was responsible for the observed latex-coagulating and -degrading activity, it was added to diluted latex and incubated at 30°C. After incubation for 24 h, controls without enzyme or with boiled enzyme were not changed, but clearing and coagulation of the latex were visible upon incubation with the active enzyme, confirming that the purified 65-kDa protein was responsible for the latex-degrading and -coagulating activity.

Isolation and identification of the main cleavage product of enzymatic rubber degradation.

In order to determine whether the purified protein cleaved the carbon backbone of the polymer or whether it only affected the integrity of the latex emulsion (i.e., the diameter-to-volume ratio of the latex particles), we investigated the presence of low-molecular-mass degradation products in the cleared and coagulated latex. Latex was incubated with the purified 65-kDa protein for several hours at 30°C and subsequently extracted with ethyl acetate; the extracts were analyzed by two-dimensional TLC. One major dark spot became visible when the TLC plate was developed with anisaldehyde, and this spot was absent in controls with either no enzyme or heat-inactivated enzyme (Fig. 2). A few minor low-intensity spots of different color also appeared in some experiments. HPLC analysis of the same ethyl acetate extract revealed only one major peak (at a retention time of 14.8 min) that was missing in control experiments (Fig. 3). The identity of the 14.8-min HPLC fraction that produced the large spot in TLC analysis was established by collecting the corresponding HPLC fraction; subsequent two-dimensional TLC analysis revealed only one spot with the same R_f value (data not shown).

FIG. 2.

Two-dimensional thin-layer chromatography of degradation products produced from latex by purified RoxA. Latex was allowed to react with purified RoxA before products were extracted with ethyl acetate, dried, and resolved with methanol. Aliquots were spotted onto TLC plates, separated with benzene-acetone (20:1) in the first dimension and with chloroform-methanol (9:1; 90°) in the second dimension, and developed with anisaldehyde-H₂SO₄ at 100°C.

FIG. 3.

Separation of polyisoprene degradation products by HPLC. Latex was incubated with purified RoxA for 3 h at 40°C and pH 7. Ethyl acetate-extracted products were loaded on a C₈ reverse-phase HPLC column as described in Materials and Methods. Heat-inactivated RoxA served as a negative control. mAU, milli absorbance units.

The compound which appeared as large dark spot in TLC and as the only prominent peak at 14.8 min in the HPLC analysis thus apparently represented the principal enzymatic rubber degradation product. In a coupled HPLC-MS analysis in the negative ESI mode, the mass spectrum of the 14.8-min fraction showed an [M-H]⁻ peak at m/z 235 (Fig. 4). When the total ion chromatogram from HPLC was analyzed for additional signals at m/z values that differed from m/z 235 by n isoprene units (i.e., Δm/z 68), additional [M-H]⁻ peaks were detected in the m/z 167, 303, 371, and 439 ion chromatograms (Table 1). The levels of these minor metabolites apparently were below the detection limit of the UV diode array in the HPLC analysis. They represent a series of homologous degradation products with one isoprene unit less or one to three isoprene units more than the major metabolite (236 Da).

FIG. 4.

HPLC-ESI-MS analysis of latex degradation products. Latex degradation products were prepared and separated by HPLC as described in the legend to Fig. 3, and the eluate was monitored by negative ESI-MS. The graph shows the average mass spectrum summed across the 14.8-min peak (seven scans). The relative intensity of the (M+1)-¹³C isotope peak (m/z 236) corresponds to a C₁₅ molecular formula.

TABLE 1.

Quasimolecular ions (M-H)⁻ of the major and minor rubber degradation products, obtained from HPLC-negative ESI-MS analysis

(M-H) m/z	No. of complete isoprene units	Retention time (min)	Relative intensity (%)a
167	1	10.95	0.2
235	2	14.82	100
303	3	18.66	11
371	4	20.73	5
439	5	22.03	0.2

^a100% = 1.62 × 10⁸ ion counts.

The positive ESI mass spectrum of the major metabolite showed an [MH]⁺ peak at m/z 237, as well as the regular adduct ions [M+Na]⁺and [M+K]⁺ and the corresponding cluster ions with methanol (m/z 259, 275, 291, and 307). Two intense peaks at m/z 219 and 201 represented the loss of one or two water molecules from the quasimolecular ion MH⁺, which definitively established the incorporation of two oxygen atoms in the metabolite (data not shown).

When the isolated 236-Da compound was reacted with dinitrophenylhydrazine, a yellow product was obtained, indicating the presence of carbonyl functions in the molecule. Additional structural information was obtained by ¹H-NMR spectroscopy of the isolated compound; the individual resonance signals are shown in Table 2 together with the first-order multiplicity and the corresponding assignments. For the resonance at the lowest field, the chemical shift (δ 9.78 ppm), relative intensity 1H, and vicinal coupling constant to the α-methylene protons (^αCH₂, 1.75 Hz) are characteristic of an aldehyde proton. The sharp singlet at 2.14 ppm (relative intensity 3H), on the other hand, is indicative of a nonconjugated acetyl group. Thus, the two ends of the main metabolite can be definitively identified as CHO-CH₂— and —CH₂-CO-CH₃, with a combined mass contribution of 100 Da. Since the overall molecular mass is 236 Da, this leaves 136 Da for the core of the metabolite, corresponding to two isoprene moieties [&0807;CH₂-C(CH₃)=CH-CH₂—; 68 Da each]. The two expected olefinic proton signals for the main metabolite (n = 2) (Fig. 5) were observed at δ 5.17 and 5.13 ppm. We concluded that 12-oxo-4,8-dimethyltrideca-4,8-diene-1-al (ODTD) (m/z 236) is the formula of the isolated degradation product with a retention time of 14.8 min in HPLC analysis. The experiments described above showed for the first time that in vitro a single enzyme is capable of cleaving the carbon backbone of rubber, yielding 12-oxo-4,8-dimethyltrideca-4,8-diene-1-al as the major degradation product. The purified protein and its corresponding gene were designated rubber oxygenase A (RoxA) and roxA, respectively.

TABLE 2.

Assignment of ¹H-NMR signals for the major metabolite (14.8-min fraction)

δ (ppm)b	Assignmenta
9.78 (t)	CHO, 3J (CHO, αCH2) 1.75 Hz
5.17 (t)	δH
5.13 (mf)	γ′H
2.49 (td), 2.44 (t)	β′H, α′H
2.35 (t), 2.24 (td)	βH, αH
2.14 (s)
2.05 (br)
1.68 (t)
1.67 (q)

^aAssignments according to the formula shown in Fig. 5A.

^bs, singlet; d, doublet; t, triplet; q, quartet; mf, multiplet with fine structure; br, broad resonance.

FIG. 5.

Proposed structure of the latex degradation products produced by RoxA. (A) Molecular structure of the main metabolite at 14.8 min in the HPLC analysis (12-oxo-4,8-dimethyltrideca-4,8-diene-1-al), including the assignments of the ¹H-NMR signals (Table 2). (B) General structure of the unequivocally characterized major and minor degradation products, with n = 1 to 5.

Biochemical characterization of the NR cleavage reaction with rubber oxygenase (RoxA).

An assay for determination of RoxA activity was developed as follows. The standard assay was performed (see Materials and Methods), and the ethyl acetate-extracted products were separated by HPLC. The RoxA activity was calculated from the peak area at 14.8 min. At low concentrations of RoxA (0.01 to 2 μg/ml of assay mixture) a linear relationship between the area of the 14.8-min peak and the RoxA concentration was found (data not shown). The optimum pH and temperature of purified RoxA were determined to be around pH 7 and 40°C, respectively (Fig. 6). The reaction was strictly dependent on the presence of oxygen, and no rubber degradation product was detected in a nitrogen atmosphere. Addition of potassium cyanide (20 mM) to the assay mixture completely inhibited the reaction. Carbon monoxide also inhibited the reaction if RoxA had been reduced by dithionite before the assay was started in the presence of oxygen (Fig. 7). The results described above are in agreement with the presence of an essential heme(s) in the enzyme. Addition of catalase at various concentrations to the assay system did not affect the reaction at all. Peroxidase activity was not detected even after a prolonged incubation time or if high concentrations of RoxA were used, while a positive control (horseradish peroxidase) reacted within seconds. Alkylation agents, such as N-ethylmaleimide and p-chlorobenzoate, and chelators (bispyridyl, tiron) partially inhibited the reaction. However, EDTA had no significant effect on the activity. Reducing agents and all of the detergents tested (except cholate) strongly inhibited the reaction (Table 3).

FIG. 6.

Optimum pH (A) and temperature (B) of purified RoxA. Latex was incubated with purified RoxA (1 μg) at different pH values or temperatures for 3 h, and ethyl acetate extracts were separated by HPLC. The area of the peak obtained at 14.8 min was used to calculate the relative amount of degradation product produced. A linear relationship between the amount of RoxA and the peak area was obtained in the range from 10 ng to 2 μg of RoxA per assay mixture. Symbols: •, 0.2 M piperazine-HCl; ○, 0.2 M bis-Tris-HCl; ▾, 0.2 M Tris-HCl; ▿, 0.2 M ethanolamine-HCl.

FIG. 7.

Dependence of RoxA activity on oxygen and effects of cyanide, carbon monoxide, and catalase. One microgram of purified RoxA was allowed to react with latex in the presence of different compounds. Ethyl acetate extracts were analyzed to determine the area of the peak at 14.8 min by HPLC. Treatments: control (air with 21% oxygen) (RoxA); catalase (1 mM); carbon monoxide (first the atmosphere was replaced by N₂ before N₂ was replaced by CO, and the assay was subsequently performed under normal air with ∼21% oxygen); air replaced by N₂; addition of 20 mM potassium cyanide.

TABLE 3.

Inhibitors of RoxA^a

Substance	Concn	Relative activity (%)
None		100
Reducing SH reagent
Dithiothreitol	5 mM	6
β-Mercaptoethanol	5 mM	<2
Alkylating reagents
N-Ethylmaleimide	10 mM	31
N-Ethylmaleimide	1 mM	47
p-Chloromercurobenzoate	10 mM	74
	1 mM	90
Chelators
α,α-Bipyridyl	10 mM	15
Tiron	10 mM	73
Tiron	1 mM	79
EDTA	10 mM	110
EDTA	5 mM	98
Detergents:
SDS	0.5%	<2
Tween 80	0.5%	<2
Sodium cholate	0.5%	32
Triton X-100	0.5%	<2
Plysurf A210G	0.5%	<2
CHAPSb	0.5%	<2
1-S-Octyl-β-d-thioglucopyranoside	0.5%	<2
Octyl-β-glucoside	0.5%	<2

^aActivities were calculated from the areas of the 14.8-min peak in the HPLC analysis (for details see the legend to Fig. 6). For inhibition of RoxA by cyanide, N₂, CO, and catalase see Fig. 7.

^bCHAPS, 3-[(3-cholamidopropyl)-dimethylammonia]-1-propanesulfonate.

RoxA appeared to be specific for oligomers and polymers of 1,4-isoprene. Natural latex and chemosynthetic poly(cis-1,4-isoprene) were significantly cleaved by purified RoxA. When a mixture of chemosynthetic rubber [oligo(cis-1,4-isoprene) with 5 to 15 isoprene units (M_w, 790; M_n, 707; M_w/M_n, 1.12)] was incubated with RoxA, a peak at 14.8 min appeared as a degradation product in HPLC, and the composition of the remaining oligomer mixture was shifted to low-molecular-weight products (data not shown). Oligomers of trans-1,4-isoprene, such as squalene, were hardly cleaved by RoxA, and only small amounts of carbonyl-containing compounds, which were slightly above the detection limit, were detected. When toluene was tested as a substrate, only background activity was obtained. Benzene did not react at all.

DISCUSSION

Biosynthetic natural rubber and chemosynthetic rubber have been used by humans at a level of about 10⁶ tons/year for about a century. The intrinsic biodegradability of polyisoprene by microorganisms in the environment has been known since the first publication of Söhngen and Fol in 1914 (16). Many other reports on biodegradation of polyisoprene have appeared during the last few decades. In particular, the number of isolated rubber-degrading microorganisms has increased during the last 10 years (7, 9, 11, 13, 17, 18). The biochemical mechanism by which the rubber backbone is cleaved is only poorly understood, however, and until now not one enzyme involved in rubber degradation has been isolated.

In this study we purified an extracellular protein from polyisoprene-grown Xanthomonas sp. cultures (RoxA) that has the ability to cleave the carbon backbone of polyisoprene in vitro, and we characterized its biochemical properties. A low-molecular-mass compound derived from a three-isoprene-unit backbone (12-oxo-4,8-dimethyltrideca-4,8-diene-1-al; m/z 236) was identified as the major degradation end product of in vitro rubber degradation by purified RoxA. Four additional minor products that differed by a mass increment of Δm/z n × 68 from the main metabolite (m/z 236) were characterized in the corresponding ion chromatograms by HPLC-negative ESI-MS analysis. Since the repetitive unit in polyisoprene has a molecular mass of 68 Da, it can be safely assumed that the minor products have the same functional groups as the main metabolite (m/z 236) and that the only structural difference is the number of isoprene units, —CH₂-C(CH₃)=CH-CH₂—, incorporated between the terminal functional groups. The concentration of these oligomers was 1 to 2 orders of magnitude lower than that of ODTD. ODTD thus apparently is the principal end product of the RoxA-catalyzed cleavage of polyisoprene. These findings are consistent with the assumption that RoxA cleaves polyisoprene oxidatively at regular intervals, cutting off three isoprene units per step. Our results are in good agreement with previous findings of Tsuchii and coworkers, who identified a whole range of related oligomers with more than 100 isoprene units in addition to ODTD (17, 18). However, these results were obtained with undefined culture fluid, and it was not known how many enzymes were involved. Presumably, the concentration and/or activity of RoxA in the culture fluid in the experiments of Tsuchii et al. was not high enough to allow complete degradation of polyisoprene.

Trypsin fingerprint analysis of RoxA confirmed that RoxA is identical to the product of a recently cloned gene assumed to be involved in rubber degradation (8). The presence of a functional signal sequence in the cloned gene was in agreement with the extracellular localization of RoxA. Comparison of the amino acid sequence of RoxA deduced from the gene with the database revealed the presence of several related amino acid sequences of hypothetical proteins. In addition to related sequences found previously (8), sequences coding for a hypothetical protein of Pirellula sp. (gi32473529), hypothetical protein Bd3821 of Bdellovibrio bacteriovorus (gi42525145), and some hypothetical proteins deduced from sequences of environmental samples were found. A function has not been identified for any of the related proteins; however, the RoxA sequence and the most closely related sequences found in the database contain a conserved sequence motif, MauG of cytochrome c peroxidases, which is consistent with the oxidative function of RoxA in Xanthomonas sp. RoxA contained approximately 2 mol of heme per mol of protein. This result is in agreement with data for the corresponding gene roxA that postulate the presence of two covalently bound heme molecules per molecule of RoxA (8). Experiments to extract heme with solvents (acid ethyl acetate or acid methyl ethyl ketone) from purified RoxA were not successful (unpublished observations), confirming the covalent binding of heme to the protein. The inhibition of RoxA by cyanide and carbon monoxide and the shift of the Soret band (406 nm) upon reduction with dithionite (418 nm) or upon incubation with synthetic rubber (409 nm) are in agreement with the involvement of heme in the reaction. Interestingly, addition of catalase did not inhibit RoxA-catalyzed cleavage of NR, suggesting that (free) hydrogen peroxide is not involved in the reaction. The negative results for RoxA in the peroxidase assay are in agreement with the latter finding. Cleavage of polyisoprene by purified RoxA was strictly dependent on the presence of molecular oxygen. In conclusion, RoxA is a novel type of oxygenase. Future experiments will address the function of heme in the reaction mechanism.