Comparison of native and recombinant chlorite dismutase from Ideonella dechloratans Helena Danielsson Thorell 1 , Natascha H. Beyer 2 , Niels H. H. Heegaard 2 , Marcus O ¨ hman 1 and Thomas Nilsson 1 1 Department of Chemistry, Karlstad University, Sweden; 2 Department of Autoimmunology, Statens Serum Institut, Copenhagen, Denmark A detailed comparison between native chlorite dismutase from Ideonella dechloratans, and the recombinant version of the protein produced in Escherichia coli, suggests the presence of a covalent modification in the native enzyme. Although t he native and r ecombinant N- and C-terminal sequences are identical, the enzymes display different electrophoretic mobilities, and produce different p eptide maps upon digestion with trypsin and s eparation of fragments using capillary electrophoresis. Comparison of MALDI m ass s pectra of tryptic peptides from the native and recombinant enzymes suggests two locations for modification in the native protein. Mass spectrometric analysis of isolated peptides from a tryptic digest of the native enzyme identifies a possible cross-linked dipeptide, suggesting an intrachain cross-link in the parent protein. Spectrophotometric titration of the native enzyme in the denatured state reveals two titrating components absorb- ing at 295 nm, suggesting the presence of about one tyrosine residue per subunit with an anomalously low pK a . The EPR spectrum for the recombinant enzyme is different from that of the native enzyme, and contains a substantial contribution of a low-spin species with the characteristics of bis-histidine coordination. These results are discussed in terms of a covalent cross-link b etween a histidine and a tyrosine sidechain, similar to those found in other heme enzymes operating under highly oxidizing conditions. Keywords: chlorate; chlorite dismutase; recombinant chlorite dismutase; post-translational modification. Chlorate- and perchlorate-respiring bacteria have attracted interest due to their potential use in the treatment of soil and water contaminated by oxyanions of c hlorine. Perchlorate, chlorate, and chlorite are recognized as potential health and environmental hazards [1–3]. In general, these compounds are not formed naturally. Rather, their appearance in the natural e nvironment is due to their manufacture and use as bleaching agents, disinfectants, h erbicides, and components of explosives and rocket propellants [4–8]. The m icrobial decomposition of oxochlorates is important in the treatment of pulp bleaching efflu ents [9], as well as in the degradation of oxochlorates released into the environment b y o ther routes [10]. Despite the fact that oxochlorates are not formed naturally, chlorate-respiring bacteria are quite ubiquitous [11,12]. Ideonella dec hloratans is a well-characterized species capable of chlorate respiration [13]. Chlorate is first converted to chlorite by a periplasmic chlorate reductase [14]. In the second step, chlorite i s decomposed to chloride and molecular oxygen by chlorite dismutase [15]. The presence of chlorite dismutase i s a prerequisite for b acterial growth as chlorite is toxic due to its high reactivity. The oxygen produced is utilized by a c ytochrome c oxidase [13]. Chlorite dismutase h as been purified, i nitially from strain GR-1 [16,17], and subsequently from strain CKB [18], and from I. dechloratans [15]. Chlorite dismutases isolated from the different species appear quite similar, being homotetra- meric heme proteins with molecular masses around 100 k Da. T he gene encoding chlorite dismutase has been cloned and sequen ced from two d ifferent species, I. dechlo- ratans [19] and Dechloromonas agitata [20]. The latter reference a lso describes a homologous gene in the genome of Magnetospirillum magnetotacticum, but in this case no expression of chlorite dismutase has been observed. The I. dechloratans chlorite dismutase gene has been expressed in Escherichia coli, and the r esulting recombinant enzyme has b een partial ly characterized [19]. In the present study, we present a more detailed characterization of recombinant chlorite dismutase, a nd a c omparison with thenativeenzyme.Ourresultssuggestthepresenceof a post-translational modification, possibly an intrachain covalent cross- link, in the enzyme p roduced in the n atural host. Materials and methods Protein purification Native chlorite dismutase was purified from I. dechloratans (ATCC 51718) a s previously described [15]. Recombinant chlorite dismutase was expressed and purified from E. coli as described i n [19], except that the cells were homogenized by a B ead-Beater (Biospec Products, B artlesville, USA). Correspondence to T. Nilsson, Karlstad University, Department of Chemistry, SE 651 88 Karlstad, Sweden. Fax: + 4 6 54 7001457, Tel.: + 46 54 7001776, E-mail: thomas.nilsson@kau.se (Received 6 May 2004, revised 8 July 2004, accepted 14 July 2004) Eur. J. Biochem. 271, 3539–3546 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04290.x Protein purity was estima ted by SDS/PAGE, and p rotein concentration w as determined by the b icinchoninic acid protein assay (Pierce Biotechnology, Rockford, IL, USA). Peptide mapping by capillary electrophoresis Recombinant and native chlorite dismutases were trans- ferred from a SDS/PAGE gel to a polyvinylidene difluoride membrane (ProBlott, Applied B iosystems, Stockholm, Sweden) by electroblotting. After staining with Ponceau S, appropriate portions were cut from the membrane and treated with trypsin (1 : 10, 37 °C, 20 h) as described in [21]. Separation of the t ryptic fragments was performed u sing a polydimethyl acrylamide coated fused silica capillary as described in [22]. Peptide mass mapping For in-gel d igestion and sample p reparation, gel plugs from SDS/PAGE stained with Coomassie brilliant blue were excised. In-gel digestion was carried out according to the protocol for silver stained bands in [23] and modified as described in [24]. Micropurification was performed accord- ing to Ku ssmann et al. [25] and Gobom et al. [26]. Samples were eluted directly onto a polished steel target plate with 0.8 lL a-cyano-4-hydroxycinnamic acid, 6 mgÆmL )1 in 0.1% trifluoro acetic a cid, 30% methanol, and 30% aceto- nitrile (premade from Agilent Technologies, Palo Alto, USA), and left to a ir-dry. For p eptide separation by RP-HPLC, the purified native enzyme was also digested by trypsin in solution. The peptides were separated b y HPLC and peak fractions were analyzed by MALDI-MS. Native chlorite dismutase (20 lL at 7 mgÆmL )1 ) was precipitated with 3 lL trichloroacetic acid (100%), left 30 min on ice, and centrifuged at 10 000 g, 15 min. The precipitate was washed with cold a cetone, vortexed and centrifuged at 10 000 g,15minandthen resuspended in 20 lLof8 M urea in 0.4 M NH 4 HCO 3 , pH 8. Water was added to a volume of 80 lL. Trypsin (4 lg) was added, and the sample was incubated with shaking at 37 °C, 52 h in a n Eppendorf Thermomixer. The digest was fractioned on a Vyd ac C18 peptide column, with a gradient of 3–97% buffer A (70% acetonitrile in 0.1% trifluoroacetic acid, v/v), 1 mLÆmin )1 , over 1 h. Fractions were collected manually, subsequently dried i n a speed-vac and resuspended in 10 lL of 0.1% (v/v) trifluoroacetic acid. One microliter was applied to the polished steel target (Scout 384) with 0.5 lL a-cyano-4-hydroxycinnamic acid (Agilent) and allowed to dry (dried droplet). Peptide mass s pectra were recorded on a Bruker UltraFlex TOF reflector mass spectrometer (Bruker Dal- tonics, Bremen, Germany), equipped with a nitrogen laser (k ¼ 337 nm). The spectra were recorded in the positive mode, using the reflector mass analyzer. Calibration was initially performed b y external c alibration using t he Bruker Peptide Standard. Whenever possible, internal mass calibration was subsequently carried out on the in-gel digestion spectra using the porcine trypsin auto- digestion products (m/z 841 .502 and 2210.096). Data analysis was c arried out by M / Z ) FREEWARE , e dition 2001.08.14 (Proteomics, N ew York, NY, USA). Database searches were carried out using PROFOUND (Proteomics), searching NCBINR , version 2002/11/27. All chemicals were analytical grade. In silico enzymatic digestion of the protein sequence and calculation of t he mass of each peptide was carried out by MS - DIGEST , ProteinProspector (http://prospector.ucsf.edu/ package). C-Terminal sequencing A C-terminal sequencer Procise 495 C (Applied Biosystems) was used for C-terminal sequencing of the native enzyme. Spectrophotometric titration Native chlorite dismutase, 6 l M (monomer), was diluted in 6 M guanidinium chloride, 10 m M borate, 10 m M Tris/HCl, pH 6. Aliquots of 1 M sodium hydroxide w ere added to the solution. At each pH value, the U V/visible spectrum was recorded using a Shimadzu U V2101 spectrophotometer. Fitting of theoretical titration curves to data was carried out using I GOR (Wavemetrics, Portland, OR, USA). Electron paramagnetic resonance (EPR) spectroscopy EPR spectra were acquired on a Bruker ER-200D-SCR spectrometer equipped with an Oxford Instruments ESR-9 helium cryostat. The concentrations of species giving rise to high- a nd low-spin signals were estimated as described in [27] and [28], respectively. Results Electrophoresis of proteins and tryptic peptides We have previously reported different electrophoretic mobilities for the native and recombinant c hlorite dismu- tases when examined by SDS/PAGE [ 19]. The recombinant enzyme migrates with a mobility close to that predicted by the amino acid sequence (corresponding to a molecular mass of 28 kDa), whereas the native enzyme migrates faster (corresponding to a m olecular mass of 25 kDa). The molecular mass, calculated from the DNA sequence, of the mature prote in is 27.8 kDa. The recombinant protein contains an extra N-terminal methionine and its predicted molecular mass is 27.9 kDa. As we have suggested [19], a possible explanation of the different mobilities is post- translational processing of chlorite dismutase in I. dechlo- ratans. Proteolytic processing at the N-terminus, however, can be excluded from the N-terminal sequencing reported in our earlier work [15]. In the present work, the C-terminal sequence was also investigated, and found to be that predicted from the gene (see below). These results exclude proteolytic processing as an explanation for the different mobilities of the native and recombinant enzymes. To investigate other covalent modifications that could affect the electrophoretic mobility, tryptic peptide maps of native and recombinant enzymes were prepared. During the course of this work we found that the recombinant enzyme was less stable than the native enzyme during the latter stages of the purification procedure, and was only possible obtain in about 70% purity. Tryptic digests were therefore prepared from proteins blotted from SDS gels to 3540 H. Danielsson Thorell et al. (Eur. J. Biochem. 271) Ó FEBS 2004 polyvinylidene difluoride membranes. The digests were analyzed by capillary e lectrophoresis using a coated capillary. F igure 1 shows electropherograms o f tryptic digests from the native and r ecombinant enzymes. Migra- tion times in this type of analysis are prone to variability [29], but most of the peptide peaks seen in the electro- pherogram of the native enzyme are also found in that obtained from the recombinant enzyme. However, there are clear differences, particularly at later migration times (marked in the figure), which are not due to migration time shifts. Thus, two peaks (denoted by arrows in Fig. 1) in the e lectropherogram of the n ative enzyme are missing in the electropherogram of the recombinant enzyme. There are also two peaks in the electropherogram of the recombinant enzyme, which do not appear to have counterparts in the native enzyme. Our finding that different peptide maps are obtained from the native and recombinant enzymes suggests a difference between their covalent structures. A lthough t he nature of such a difference cannot be inferred from these results, we note that anomalously high electrophoretic mobilities in SDS/ PAGE analyses have been observed in p roteins containing covalent cross-links, such as disulfide bonds [30,31] or cross-links caused by oxidative c oupling of sidechains [32,33]. The higher electrophoretic mobility in these proteins is probably due to the smaller hydrodynamic radius caused by the cross-link. Mass spectrometry Detailed investigations of possible differences between native and recombinant chlorite dismutase covalent structure were carried out using MALDI-TOF mass spectrometry. Tryptic peptide mass maps of the native enzyme, from in-gel digestion and digestion in solution, were analyzed. Masses covering most of the predicted amino acid sequence of t he enzyme could be i dentified in these s pectra, when a llowing four missed cleavages in the tryptic in silico digestion. The sequence coverage based on the mass spectra, and on C-terminal sequencing of the n ative enzyme, is shown in Fig. 2A. Four fragments, corresponding to HK(52–53), RK(180–181), VPENKYHVR(215–223) and T(242) (bold) were not covered. To compare the native and recombinant enzymes, peptides were generated by in-gel trypsin digestion and subject to mass analysis using as above. For the native enzyme, we obtained basically the s ame sequence coverage as above. The recombinant enzyme produced, however, a prominent peak at a mass of 1571.7, which is completely absent in the native enzyme. A comparison of the mass spectra obtained from the native and r ecombinant enzymes is shown in Fig. 3. Analysis of the sequence reveals the fragment HKEKVIVDAYLTR(52–64) (Fig. 2B) as the probable origin of t his p eak. This fragment i ncludes HK(52–53), which is missing in the sequence coverage of the native enzyme. This result implicates HK(52–53) as a possible l ocation for a covalent modification. The fragment VPENKYHVR(215–223), also missing in the m ass spectra, is another possible location. In the mass spectrum of recombinant enzyme, VPENK(215–219) was absent, whereas YHVR(220–223) was observed as a part of fragment (220–241). To identify modified fragments, t ryptic peptides from the native enzyme were separated by HPLC and individually analyzed by MS. Matching sequence coverage was obtained after a nalysis of the mass spectra of the individual peptide fractions. One peptide fraction from t he chromatographic separation produced a mass spectrum containing a peak (m/z ¼ 1679.8) (Fig. 4), corresponding to the sum (minus one hydrogen) of the fragments containing HKEK(52–55) and VPENKYHVR(215–223) (Fig. 2C). We could not, however, detect a fragment at m/z ¼ 1426 corresponding to fragment (52–53) combined wit h fragment (215–223). Localization of a modification to fragment (52–53) is therefore t entative. Fig. 1. Separation of tryptic peptides of native (A) and re combinant (B) c hlorite dismutase by capillary e lectrophoresis with a polydimethyl acrylamide-coated fused silica capillary. Dashed arrows indicate correspondence, and solid arrows denote peaks that d o not h ave counterparts in the other e lectro pherogram. Ó FEBS 2004 Native and recombinant chlorite dismutase (Eur. J. Biochem. 271) 3541 Spectrophotometric titration of the tyrosines in native chlorite dismutase Several of the covalent cross-links observed earlier [ 34] include tyrosine sidechains, and we note the presence of tyrosine in the VPENKYHVR fragment (215–223) that was indicated by t he mass spectrometry analyses to be involved in a cross-link. Cappuccio et al. [35] and McCauley et al. [36] showed, using spectrophotometric titrations of model compounds, t hat a cross-linked t yrosine has lowered pK a value of the phenolic proton. To investigate the possibility of tyrosine sidechains with an anomalously low pK a value i n chlorite dismutase, spectrophotometric t itration of the native enzyme, completely denatured in 6 M guanidinium chloride, was carried out. Figure 5 shows the absorbance at 295 n m (the absorption m aximum of t he tyrosinate ion [ 37] as a function of pH. A curve fit of a single titration curve (Fig. 5 A) did not yield a satisfactory fit, suggesting the presence of more than one titrating component. This is not expected when the enzyme is completely denatured, as all tyrosines should b e in t he same chemical environment. A curve fit with two titrating components gave a better fit (Fig. 5 B). The major component, accounting for 92% of the total amplitude, t itrated with a pK a value of 1 0.15 ± 0.0 3, in accordance with the pK a value of 10.1 for tyrosine [35]. For the minor component, accounting for 8% of the total amplitude, a pK a value o f 8 .35 ± 0.3 w as found. This is similar to the value found for the histidine methyl ester derivative studied in [35]. Chlorite dismutase contains 12 tyrosine residues per subunit. We note that the fraction of the minor component corresponds to about one of the 12 tyrosines titrating with the lower pKa value. EPR The EPR spectrum o f the recombinant c hlorite dismutase at pH 7 is shown in Fig. 6. In contrast to the EPR spectrum o f the native enzyme at neutral pH (trace A; see Fig. 3. Mass analyses of tryptic peptides from native (A) and recombinant (B) chlorite dismutase. Only the 1558–1615 mass range is shown. Fig. 2. Sequences for the complete p rotein and for detected fragments of chlorite dismutase. (A) The native chlorite dismutase amino acid sequence with the coverage obtained by u sing MALD I-MS. The bold and italic sequences were not detected. The sequence in it alics i s t hat obtained in C-terminal sequencing. (B) The calculated monoisotopic mass [MH] + of the peptides a re shown. (C) The monoisotopic size of the peptides tha t would result from h istidine–tyrosine covalent linkage (1679.92 D a). 3542 H. Danielsson Thorell et al. (Eur. J. Biochem. 271) Ó FEBS 2004 also [15]), which contains only high-spin heme, the spectrum of the recombinant enzyme (trace B) is a mixture of contributions from high- and low-spin species. The high-spin heme component in the spectrum consists of both a rhombic and an axial species with a total concentration of 58 l M . The majority of the high-spin heme has the characteristics of a rhombically distorted heme. From th e spectrum, the g-values 6.31 and 5.47 are obtained. Essentially the same g-values are found in the EPR spectrum o f native chlorite dismutase at neutral pH. A minor part of the high-spin heme is axial with a g-value at 5.9. This axial high-spin heme is not found in the spectrum of the native enzyme from I. dechloratans but a similar component was observed the in EPR spectra of chlorite dismutase from strain GR-1 recorded at neutral pH [38]. For the l ow-spin component, the, g-values at 3.04, 2.25, and 1.52 are obtained. The integrated ampli- tude for this signal corresponds to a concentration of 43 l M , which is little less than half of the total heme concentration. Fig. 4. MALDI mass spectrum of at HPLC fraction of tryptic digest of nativ e chlorite dismutase. The 1679.8 Da mass fra gment is denoted by an arrow. The inset shows an expanded view of the 1600–1700 m ass range. Fig. 5. Spectrophotometric titration of tyrosine residues in the denatured na tive chlorite dismu- tase. The titration was monito red at 295 nm at which only tyrosinate absorbs. The protein contains 1 2 tyrosine residues per s ubunit. (A) The solid line is the result of curve fitting with A tot ¼ 0.216 and pK a ¼ 10.1. (B) The solid line is the result of curve fitting with A tot1 ¼ 0.206, pK a1 ¼ 10.15, A tot2 ¼ 0.017, pK a2 ¼ 8.35. Ó FEBS 2004 Native and recombinant chlorite dismutase (Eur. J. Biochem. 271) 3543 Discussion The detailed characterization of recombinant I. dechlora- tans chlorite dismutase, and comparison with the native enzyme carried out here, suggest the presence of a covalent modification in chlorite dismutase produced in the n atural host, but not in the recombinant version of the enzyme. Comparison of mass spectra for tryptic peptides obtained from the native a nd recombinant e nzymes suggest HK( 52– 53) and YHVR(220–223) as sites of modification. Further- more, a fragment, i solated by HPLC, in the tryptic digest of the native enzyme could be identifi ed as a possible product of cross-linking between HKEK(52–55) and VPENKYHVR(215–223) (Fig. 2C). Cross-linking is an attractive candidate for covalent modification, as it would account also for the higher electrophoretic mobility (due to the smaller hydrodynamic radius) observed for the native enzyme, and for the different peptide maps observed after tryptic cleavage and separation by capillary electrophoresis. The nonenzymatic formation of covalently or oxida- tively modified amino acids has been demonstrated [39–41], an d s everal cases of c ross-links inc luding histidine and tyrosine residues in oxidative enzymes have been reported recently [34]. The crystal structure of galactose oxidase revealed that t he enzyme contained a modified active site tyrosine covalently cross-linked to a cysteine at the ortho-position to t he phenolic oxygen [42]. M ore recently, cytochrome c oxidase has also been found to contain a modified tyrosine, with the crystal structures showing a covalent link between the active site tyrosine (at the ortho-position) to the imidazole N e of a histidine [43,44]. A different type of histidine–tyrosine cross-link was discovered in the crystal structure of catalase HPII from E. coli [45,46]. In t his case, a covalent bond joins C b of the essential t yrosine and one of the imidazole nitrogens (N e ) of a histidine. These observations, together with the presence of a t yrosine residue in the fragment suggested to contain the cross-link (Fig. 2C), prompted us to investi- gate the presence of m odified tyrosines. The results obtained from spectrophotometric titration of the native enzyme denatured in g uanidinium chloride suggest that about one of the 1 2 tyrosines in chlorite d ismutase t itrates with an anomalously low p K a value. The pK a value obtained (8.3) is similar to that found for a model histidine-phenol compound, 1-o-phenol(acetyl)histidine methyl ester [35], and suggests the presence of a modified tyrosine in chlorite dismutase. The result of the spectro- photometric titration, t ogether with the mass spectrometric data implicating the tyrosine-containing fragment VPEN- KYHVR(215–223) as a p art of a cross-link, is consistent the participation of tyrosine in cross-linking. From the low pK a value of 8.3 found in the spectrophotometric titration, the catalase HPII variant of cross-link is less likely as substitution at the C b is not expected to affect the phenolic pK a value. An histidine–tyrosine bond may be somewhat labile [45] and t his, in addition to ionic suppression, could explain the rather low y ield of the dipeptide fragment mass in the MS analyses. T he fragmented dipeptide would not necessarily yield its const ituent two try ptic frag ment p eptide ma sses as fragmentation may involve various parts of t he molecule and sidechains may be derivatized. Also, the small mole- cular mass part of the dipeptide would be prone to be obscured in the area of the mass spectrum dominated by signals from matrix components. The environment of t he heme group in the recombinant enzyme was investigated using EPR s pectroscopy. In contrast to what is observed i n the native enzyme, the EPR spectrum sho ws the p resence of s everal species. The major components are a high-spin species with a spectrum similar to that observed in the native enzyme, a nd a low- spin species. An earlier characteriz ation using optical spectroscopy [19] also revealed two components, one with a native-like spectrum and one with absorption maxima at 405 and 525 nm in the oxidized state. The later species could not be reduced by dithionite. T his species is probably the same as the one displaying the low-spin EPR signal. The g-values of t his component a re different from those found for the low-spin component in the EPR spectrum of native chlorite dismutase at high pH (2.56, 2.19, and 1.87) [15], and they are also distinct f rom those found in other hydroxide-coordinated systems [47]. The g-values are more similar t o those observed f or bis- histidine coordinated heme [47]. Moreover, a similar EPR spectrum was observed i n [38] after addition of imidazole to chlorite dismutase from GR-1. Therefore, the heme group is p robably coordinated by two histidine s idechains in the low-spin component of the recombinant chlorite dismutase. These results suggest a difference in structure of the heme pocket in the native and recombinant e nzymes, with a histidine sidechain being more accessible for heme coordination from the distal side in the recombinant enzyme. The difference between the heme environmen ts in the native and recombinant enzymes is probably due to structural differences caused b y covalent cross-linking. Fig. 6. EPR spectra of native and of recombinant chlorite d ismutase at neutral pH. (A) Native chlorite dismutase. (B) Recombinant chlorite dismutase. Prote in concentrations we re about 100 l M (hem). EPR conditions: temperature 10 K; microwave power, 2 mW; mic rowave frequency, 9.449 GHz; modulation amplitude, 20 G. 3544 H. Danielsson Thorell et al. (Eur. J. Biochem. 271) Ó FEBS 2004 As discussed above, the histidine residue in fragment (52– 53) could b e involved in cross-linking, and an interestin g possibility is that t his residue is available for coordination in the recombinant enzyme where a cross-link is not present. Cross-links involving oxidatively coupled sidechains have been found in enzymes operating under highly oxidizing condition, and have been suggested to originate from radicals formed in the reaction of the heme group with oxidants. In cytochrome c oxidase, a tyrosyl radical is formed during the reaction of the mixed-valence state of the enzyme with oxygen [ 48]. MacMillan et al.[49]have reported the EPR signal of a radical generated in cyto- chrome c oxidase. This signal was proposed to originate from the cross-linked tyrosine. In catalase, the compound I, containing Fe(IV) and a porphyrin radical is produced after the r eaction with one equivalent of hydrogen peroxide. For catalase HPII, which contains a histidine–tyrosine cross- link, it has been proposed that compound I is the species in which the post-translational modification takes place [45,50,51]. Although the catalytic mechanism of chlorite dismutase is not known, the formation o f similar interme- diates appears likely, given t he nature of the reactant. The reaction of chlorite with other heme enzymes, horseradish peroxidase and chloroperoxidase, has been shown to produce the highly oxidized compound I [52]. M oreover, a radical signal is present in the EPR s pectrum of chlorite dismutase f rom strain GR-1 [38]. The formation of a c ross- link i n chlorite dismutase by oxidative coupling, similar to the mechanisms suggested for cytochrome c oxidase [41,48,49] and catalase HPII [45,50,51], therefore appears possible. Cross-linking is expected to increase the stability of a protein, and it absence in the recombinant enzyme could account for the lower stability during t he latter stages of its purification. The catalytic properties of the recombinant enzyme are, however, similar t o those of t he native enzyme, suggesting cross-linking is not important for catalysis. This would be similar to cytoch rome c oxidase, where the histidine–tyrosine cross-link has been suggested to play role in preserving the binuclear site architecture [40,41,53,54]. In conclusion, our comparison between the n ative and recombinant I. dechloratans chlorite dis mutase suggests that the enzyme produced in the natural host contains a covalent modification, probably an intrachain cross-link involving a residue in the 52–55 region and a residue in the 215–223 region. A tyrosine–histidine cross-link appears possible, and could account for EPR differences between the native and recombinant enz ymes as well as the spectrophotometric titration of the native enzyme. More work is, however, needed to establish the nature of the m odification. Acknowledgements We thank Roland Aasa ( Chalmers University of Technology, Sweden) for recording the EPR spectrum and for h elpful suggestions regarding its interpretation. We also thank Annika Norin and Ella Cederlun d (Karolinska i nstitutet, Sweden) for C-terminal amino a cid sequencing of the n ative enz yme, and Justyna M. Czarna for help with the mass spectrometric analyses. References 1. Rosemarin, A., Mattson, J., Lehtinen, K J., Notini, M. & Nyle ´ n, E. (1986) Effects of pulp mill chlorate on Fucus vesiculosus – a summar y of projects. Ophelia Suppl. 4, 219–224. 2. Urbansky, E.T. (1998) Perchlorate chemistry: implications for analysis and remediation. Bioremediation J. 2, 81–95. 3. Renner, R. (2003) Environmental health: academy to mediate debate over ro cket-fu el contaminants. Science 299, 1829. 4. A ˚ slander, A. (1928) Experiments on the eradication of canada thistle, Cirsum arve nse, with c hlorates and other herbicides. J. Agric. Res. 36, 915–934. 5. Germga ˚ rd, U., Teder, A. & Tormund, D. (1981) Chlorate for- mation during chlorine dioxid e bleac hing of softwoo d kraft pu lp. Pap. Puu 63, 1 27–133. 6. Rosemarin, A., Lehtinen, K J., Notini, M . & Mattson, J. (1994) Effects of pulp mill chlorate on baltic sea algae. Environ. Pollut. 85 , 3–13. 7. Herman, D.C. & Frankenberger, W.T.J. (1999) Bacterial reduc- tion of p erchlorate a nd nitrate i n w at er. J. Environ. Qual. 28, 1018– 1024. 8. Hogue, C. (2003) Rocket-fueled river. Chem. Eng. News 81, 37–46. 9. van Wijk, D.J., Kroon, S.G.M. & Garttener-Arends, I.C.M. (1998) Toxicity of chlorate and chlorite to selected species of algae, bacteria, and fungi. Ecotoxicol. Environ. Safety 40 , 206–211. 10. Logan, B.E. (1998) A review of chlorate- and perchlorate-respiring microorganisms. Bioremediation J. 2, 69–79. 11. O’Connor, S.M. & Coates, J.D. (2002) Universal immunoprobe for (per)chlo rate-reducing b acteria. Appl. Environ. Microbiol. 68, 3108–3113. 12. Lovley, D.R. & Coates, J.D. ( 2000) Novel forms of anaerobic respiration of environmental relevance. Curr. Opin. Microbiol. 3, 252–256. 13. Malmqvist, A ˚ ., Welander, T., Moore, E., Ternstro ¨ m, A., Molin, G. & Stenstro ¨ m, I. (1994) Ideonella dechloratans Generalnov., sp.nov., a new bacterium capable of growing anaerobically with chlorate as an electron acceptor. System. Appl. Microbiol. 17, 58–64. 14. DanielssonThorell,H.,Stenklo,K.,Karlsson,J.&Nilsson,T. (2003) A gene cluster for chlorate metabolism in Ideone lla dechloratans. Appl. E nviron. Microbiol. 69, 5585–5592. 15. Stenklo, K., Danielsson T horell, H., B ergius, H., Aasa, R . & Nilsson, T. (2001) Chlorite dismutase from Ideonella dechloratans. J. Biol. Inorg. Chem. 6, 601–607. 16. Rikken, G.B., Kroon, A.G. & van Ginkel, C.G. (1996) Trans- formation o f (per)chlorate into chloride by a newly isolated bac- terium: red uction and dismutation. Appl. Microbiol. Biotechnol. 45, 420–426. 17. Kengen, S.W., Rikken, G.B., Hagen, W.R., van Ginkel, C.G. & Stams, A.J. (1999) Purification and characterization of (per)- chlorate reductase from the chlorate-respiring strain GR-1. J. Bacteriol. 181 , 6706–6711. 18. Coates, J.D., Michaelidou, U., Bruce, R.A., O’Connor, S.M., Crespi, J.N. & Achenbach, L.A. (1999) Ubiquity a nd diversity of dissimilatory (per)chlorate-reducing bacteria. Appl. Environ. Microbiol. 65, 5234–5241. 19. Danielsson Thorell, H., Karlsson, J., Portelius, E. & Nilsson, T. (2002) Cloning, characterisation, and expression of a novel gene encoding chlorite dismutase from Ideonella dechloratans. Biochim. Biophys. Acta 1577, 4 45–451. 20. Bender, K .S., O’Connor, S .M., Chakraborty, R ., Coates, J .D. & Achenbach, L.A. (2002) Sequencing and transcriptional analysis of the chlorite dismutase gene of Dechloromonas agitata and its u se as a metabolic probe. Appl. Environ. Microbiol. 68, 4820–4826. 21. Walker, J.M. (1998) Protein Protocols on CD-ROM.Humana Press Inc., Totowa, N J, USA. Ó FEBS 2004 Native and recombinant chlorite dismutase (Eur. J. Biochem. 271) 3545 22. Wan, H., O ¨ hman, M. & Blomberg, L.G. (2001) Bonded dimethylacrylamide as a p ermanent coating for capillary electro- phoresis. J. Chr omatogr. A 924, 59–70. 23. Shevchenko,A.,Wilm,M.,Vorm,O.&Mann,M.(1996)Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68, 850 –858. 24. Jensen, O.N., Larsen, M.R. & Roepsto rff, P. (1998) Mass spec- trometric identification and microcharacterization of proteins from electrophoretic gels: strategies and applications. Proteins Suppl. 2, 74–89. 25. Kussmann, M., Lassing, U., Sturmer, C.A., Przybylski, M. & Roepstorff, P. (1997) M atrix-assisted laser desorption/ionization mass spectrometric peptide mapping of th e neural cell adhesion protein n eurolin purified by sodium dodecyl sulfate polyacryl- amide gel electrophoresis or acid ic pre cipitation. J. Mass Spec- trom. 32, 4 83–493. 26. Gobom, J., Nordhoff, E., Mirgorodskaya, E., Ekman, R. & Roepstorff, P. (1999) Sample purification a nd preparation tech- nique b ased on nano-scale reverse d-phase columns for the sensi- tive analysis of complex peptide mixtures by matrix-assisted laser desorption/ionization mass spectrometry. J. Mass Spe ctrom. 34, 105–116. 27. Aasa, R., Albracht, P.J., Falk, K .E., Lanne, B. & Va ¨ nnga ˚ rd, T. (1976) EPR signals from cytochrome c oxidase. Biochim. Biophys. Acta 422, 2 60–272. 28. Aasa, R. & Va ¨ nnga ˚ rd,T.(1975)EPRsignalandintensityand powder shapes: a reexamination. J. M agn. Reson. 19, 308–315. 29. Grossman, P.D. & Colburn, J.C. (1992) Capillary Electrophoresis. Academic Press, Inc, San D iego, CA. 30. Selimova, L.M., Zaides, V.M. & Zhdanov, V.M. (1982) Disulfide bonding in influenza virus proteins as revealed by polyacrylamide gel electrophoresis. J. Virol. 44, 450–457. 31. Shvetsov, A., Musib, R., Phillips, M., Rubenstein, P.A. & Reisler, E. (2002) Locking the hydrophobic l oop 262–274 to G-actin sur- face by a disulfide bridge prevents filament formation. Biochem- istry 41, 10787–10793. 32. Baron, A.J., Stevens, C., Wilmot, C., Seneviratne, K.D., Blakeley, V., Dooley, D.M., Phillips, S.E., K nowles, P.F. & McPherson, M.J. (1994) Structure and mechanism o f galactose oxidase: the free radical site. J. Biol. Chem. 269, 25095–25105. 33. Whittaker, M.M. & Whittaker, J.W. (2003) Cu(I)-dependent biogenesis of the galactose oxidase redox cofactor. Biol. Chem. 278, 22090–220101. 34. Okeley,N.M.&vanderDonk,W.A.(2000)Novelcofactorsvia post-translational modifications of enzyme active sites. Chem. Biol. 7, R159–R171. 35. Cappuccio, J.A., Ayala, I., Elliott, G.I., Szundi, I., Lewis, J., Konopelski, J .P., Barry, B.A. & Einarsdottir, O. (2002) Modeling the active s ite of cytochrome o xidase: synthesis a nd characteriza- tion of a cross-linked histidine-p henol. J. Am. Chem. Soc. 124, 1750–1760. 36. McCauley,K.M.,Vrtis,J.M.,Dupont,J.&vanderDonk,W.A. (2000) Insights into the functional role o f t he tyrosine-histidin e linkage in cytochrome c oxidase. J. Am. C hem. Soc. 122, 2403– 2404. 37. Tanford, C., H auenstein, J.D. & Rands, D .G. (1956) Phenolic hydroxyl ionization in proteins II ribonuclease. J. Am. Chem. Soc. 77, 6409–6410. 38. Hagedoorn, P.L., D e G eus, D.C. & Hagen, W.R. ( 2002) S pec- troscopic ch aracterization a nd ligand-bindin g p roperties of chlorite dismutase from the chlorate respiring bacterial strain GR-1. Eur. J. Biochem. 269, 4905–4911. 39. Dooley, D.M. (1999) Structure and biogenesis of topaquinone and related cofactors. J. Biol. Inorg. Chem. 4, 1–11. 40. Rogers, M.S. & Dooley, D.M. (2001) Posttranslationally modified tyrosines f rom galactose oxidase and c ytochrome c o xidase. Adv. Protein Chem. 58, 387–436. 41. Rogers, M.S. & Dooley, D.M. (2003) Copper-tyrosyl radical enzymes. Curr. Opin. Chem. Biol. 7, 189–196. 42. Ito, N., P hillips, S.E., St evens, C., Ogel, Z.B., M cPherson, M.J., Keen, J.N., Yadav, K.D. & Knowles, P.F. (1991) Novel thioether bond revealed by a 1.7 A ˚ crystal structure of galactose oxidase. Nature 350, 87–90. 43. Ostermeier, C ., Harrenga, A., Ermler, U. & Michel, H. (1997) Structure at 2.7 A ˚ resolution of the Paracoccus denitrificans two- subunit cytoc hrome c oxidase c omplexed with an antib ody FV fragment. Proc. Natl Acad. Sci. USA 94, 10547–10553. 44. Yoshikawa, S., S hinzawa-Itoh, K ., Nakash ima, R., Yaono, R., Yamashita, E., Inoue, N., Yao, M., Fei, M.J., Libeu, C.P., Mizushima, T., Yamaguchi, H., Tomizaki, T. & Tsukihara, T. (1998) Redox-c ouple d cryst al stru ctural ch ang es in bovine hear t cytochrome c oxidase. Science 280, 1723–1729. 45. Bravo, J., Fita, I., Ferrer, J.C., Ens, W., Hillar, A. , Switala, J. & Loewen, P.C. (1997) Identification of a novel bond b etween a histidine and the essential tyrosine in catalase HPII of Escherichia coli. Protein Sci. 6, 1016–1023. 46. Bravo, J., Mate, M.J., Schneider, T., Switala , J., Wilson, K., Loewen, P.C. & Fita, I. (1999) Struc ture of catalase HPII from Escherichia coli at 1 .9 A ˚ resolution. Proteins 34 , 155–166. 47. Gadsby, P.M.A. & Thomson, A.J. (1990) Assignment of the axial ligands of ferric io n in low-sp in h emoproteins by near-infrared magnetic circular dichroism and electron paramagnetic resonance spectroscopy. J. Am. Chem. Soc. 112, 5003–5011. 48. Proshlyakov, D.A., Pressler, M.A. & Babcock, G.T. (1998) Dioxygen activation and bond cleavage by m ixed-valence cyto- chrome c oxidase. Proc. N atl Acad. Sci. USA 95 , 8020–8025. 49. MacMillan, F., Kannt, A ., Behr, J., Prisner, T. & M ichel, H. (1999) Direct evidence for a tyrosine radical in the reaction of cytochrome c oxidase with hydrogen peroxide. Biochemistry 38, 9179–9184. 50. Mate, M.J., Sevinc, M.S., Hu, B., Bujons, J., Bravo, J., Switala, J., Ens, W., Loewen, P.C. & Fita, I. (1999 ) Mutants that alter the covalent structure of c atalase hydroperoxidase II from Escherichia coli. J. Biol. C hem. 274, 2 7717–27725. 51. Melik-Adamyan,W .,Bravo,J., Carpena,X.,Switala,J .,Mate,M.J., Fita,I.&Loewen,P.C.(2001)Substrate flow in catalases deduced from the crystal structures of active site variants o f HPII from Escherichia coli. Proteins 44, 270–281. 52. Hollenberg, P.F., Rand-Meir, T. & Hager, L.P. (1974) The r eac- tion of chlorite with horseradish peroxidase and chloroperoxidase: enzymatic chlorination and spectral intermediates. J. Biol. C hem. 249, 5816–5825. 53. Das, T.K ., Pec oraro, C., Tomson, F.L., Gennis, R.B. & Rousseau, D.L. ( 1998) The post-translational modification in cyto chrome c oxidase is required to establish a functional environme nt of the catalytic site. Biochemistry 37, 14471–14476. 54. Pinakoulaki,E.,Pfitzner,U.,Ludwig,B.&Varotsis,C.(2002) The role o f the cross-link H is-Tyr in the functional properties of the binuclear center in cytochrome c oxidase. J. Biol. Chem. 277, 13563–13568. 3546 H. Danielsson Thorell et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . 6. EPR spectra of native and of recombinant chlorite d ismutase at neutral pH. (A) Native chlorite dismutase. (B) Recombinant chlorite dismutase. Prote. Copenhagen, Denmark A detailed comparison between native chlorite dismutase from Ideonella dechloratans, and the recombinant version of the protein produced in