Cloning of the manganese lipoxygenase gene reveals homology with the lipoxygenase gene family Lena Ho¨ rnsten 1 , Chao Su 1 , Anne E. Osbourn 2 , Ulf Hellman 3 and Ernst H. Oliw 1 1 Department of Pharmaceutical Biosciences, Uppsala Biomedical Centre, Uppsala, Sweden; 2 The Sainsbury Laboratory, John Innes Centre, Norwich, UK; 3 Ludwig Institute for Cancer Research, Uppsala Biomedical Centre, Uppsala, Sweden Manganese lipoxygenase was isolated to homogeneity from the take-all fungus, Gaeumannomyces graminis. The C-ter- minal amino acids and several internal peptides were sequenced, and the information was used to obtain a cDNA probe by RT/PCR. Screening of a genomic library of G. graminis yielded a full-length clone of the Mn-Lipoxyg- enase gene. cDNA analysis showed that the gene spanned 2.6 kb and contained one intron (133 bp). Northern blot analyses indicated two transcripts (2.7 and 3.1 kb). The deduced amino-acid sequence of the Mn-Lipoxygenase precursor (618 amino acids, 67.7 kDa) could be aligned with mammalian and plant lipoxygenases with 23–28% identity over 350–400 amino-acid residues of the catalytic domains. Lipoxygenases have one water molecule and five amino acids as Fe ligands. These are two histidine residues in the highly conserved 30 amino-acid sequence WLLAK-X 15 - H-X 4 -H-X 3 -E of a helix 9, one histidine and usually an asparaine residue in the sequence H-X 3 -N-X-G of a helix 18, and the carboxyl oxygen of the C-terminal isoleucine (or valine) residue. The homologous sequence of a helix 9 of Mn-Lipoxygenase [WLLAK-X 14 -H(294)-X 3 -H(297)-X 3 -E] contained two single-amino-acid gaps, but otherwise His294 and His297 aligned with the two His residues, which coordinate iron. Mn-Lipoxygenase [H(478)-X 3 -N(482)-X-G] could be aligned with the two metal ligands of ahelix 18, and the C-terminal residue was Val618. We conclude that Mn-Lipoxygenase belongs to the lipoxygenase gene family and that its unique biochemical properties might be related to structural differences in the metal centre and a helix 9 of lipoxygenases rather than to the metal ligands. Keywords: ascomycete; dioxygenase; lipoxygenase; hydro- peroxide; metalloenzyme. Lipoxygenases (LOX; EC 1.13.11.12) are widely distributed in mammals and plants and oxygenate polyunsaturated fatty acids to cis–trans conjugated hydroperoxides [1]. LOX have three important biological functions. The hydroperoxy fatty acids may act as signal molecules, either directly or after conversion to a large variety of biologically active products such as leukotrienes in man [2] and jasmonic acid in plants [3]. LOX can also catalyze physiological break- down of cellular membranes and organelles in the lens and in the reticulocyte [1,4]. Plant LOX genes are activated in response to wounding and pathogen attack [5], and reduced plant LOX activity results in an increased susceptibility to insects and fungal pathogens [6,7]. All LOX belong to the same gene family [1]. The pair- wise amino-acid sequence identity of plant and animal LOX is only 21–27%, whereas the corresponding figures within pairs of plant or pairs of animal LOX are often 40% or higher. LOX in animals and plants contain mononuclear nonheme Fe as the catalytic metal, which has been demonstrated by atomic absorption spectroscopy for soybean LOX [8], rabbit reticulocyte 15-LOX [9] and human 5-LOX [10]. X-ray crystallography of soybean LOX L1 and L3 [11–14], and rabbit reticulocyte 15-LOX [15] has identified the Fe(II) ligands. These are one water molecule and five amino acids [12]. The iron ligands are the carboxyl oxygen of the C-terminal isoleucine (or valine) residue, the nitrogen atoms of two histidine residues of a helix 9 and one histidine residue of a helix 18, and the distant amid oxygen of an asparagine residue of a helix 18 [1,12,16]. There is a large group of nonheme Fe(II) enzymes, which have a common struc- tural motif, the 2-His-1-carboxylate facial triad. This triad designates two histidine nitrogens and the carboxyl oxygen of asparagine, glutamic acid, or the C-terminal isoleucine residue as three of the Fe ligands, and LOX is considered to belong to this group of enzymes, although LOX has five metal ligands [17]. Early reports suggest that LOX occur in fungi, but the enzymes have not been described in detail (for a review see [18]). The take-all fungus, Gaeumannomyces graminis, which is a root pathogen of wheat, forms the only fungal LOX that has been characterized. This LOX has several unique properties [19]. First, it contains Mn in its active center and it was therefore designated Mn-LOX. Manganese is tightly bound to the apoenzyme in a 1 : 1 stoichiometry, and cannot be extracted with metal chela- tors. Second, the enzyme metabolizes linoleic and a-linolenic acids to 13R-hydroperoxy fatty acids and to novel LOX products, 11S-hydroperoxy fatty acids [20]. Correspondence to E. H. Oliw, Division of Biochemical Pharmacology, Department of Pharmaceutical Biosciences, Uppsala University, PO Box 591, Husargatan 3, SE-751 24 Uppsala, Sweden. Fax: + 46 18 55 29 36, Tel.: + 46 18 471 44 55, E-mail: Ernst.Oliw@farmbio.uu.se Abbreviations: LOX, lipoxygenase(s); Gga, G. graminis var avenae; Ggt, G. graminis var tritici; Mn-LOX, manganese lipoxygenase. Enzymes: lipoxygenases (EC 1.13.11.12). Note: The sequences reported in this paper have been deposited in GenBank under accession nos AY040824 and AY040825. (Received 4 March 2002, accepted 17 April 2002) Eur. J. Biochem. 269, 2690–2697 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02936.x Third, Mn-LOX is the first LOX known to be secreted by a microorganism, and it is also remarkably stable [19]. The biological function of Mn-LOX is unknown, but the enzyme may cause oxidative damage and contribute to the pathogenicity of G. graminis. Analysis of the metal cofactor of Mn-LOX during catalysis revealed important similarities with LOX. The mononuclear metal center of Mn-LOX redox cycles between Mn(II) in the resting state and Mn(III) in the active state [21], whereas the metal centre of LOX redox cycles in the same way between Fe(II) and Fe(III) [1]. The active forms of both enzymes abstract, with stereo-speci- ficity, a bisallylic hydrogen from their fatty acid substrates and form a substrate radical. The free radical reacts with molecular oxygen in a controlled fashion relative to the hydrogen abstraction so that antarafacial oxygen insertion is catalyzed by LOX and suprafacial oxygen insertion by Mn-LOX [1,20]. The metal ligands contribute to the large diversity of nonheme Fe(II) enzymes [17]. Some enzymes occur in homologous forms with Fe or Mn as catalytic metals, and the metal ligands can be conserved. The extradiol-cleaving catecholdioxygenase(3,4-dihydroxyphenylacetate2,3-dioxy- genase) occurs in two homologous forms with either prosthetic Fe or Mn [22]. X-ray crystallography of the Fe form of the 2,3-dioxygenase shows that Fe is coordinated to three amino-acid residues (His145, His209 and Glu260) and to two molecules of water [23,24]. Site-directed mutagenesis of the Mn form suggests that the correspond- ing conserved amino acids (His155, His214 and Glu266) are essential for catalytic activity [25]. Superoxide dismu- tases also have identical metal ligands and tertiary fold for Fe- and Mn-dependent forms [26]. This left us with the intriguing possibility that Mn-LOX and LOX might have identical metal ligands and yet form different oxidation products. The aim of the present investigation was to clone and sequence the Mn-LOX gene. EXPERIMENTAL PROCEDURES Materials [a- 32 P]dCTP (3000 CiÆmmol )1 ), dNTPs, [a- 33 P]ddNTPs, Hybond-N membranes, DNA labeling beads (dCTP), and T-primed first-strand kit were from Amersham Pharmacia Biotech. TA cloning kits were from Invitrogen. Taq DNA polymerase and the enhanced avian RT/PCR kit were from Sigma. Restriction enzymes were from New England BioLabs. Two strains of G. graminis [var avenae (Gga) and var tritici (Ggt)] were obtained and grown as described [19,27,28]. Qiagen plant DNeasy mini, RNeasy mini and QIAquick gel extraction kits were from Merck Eurolab (Stockholm, Sweden). Degenerate primers for PCR were obtained from TIB Molbiol (Berlin, Germany), and sequencing primers were from CyberGene (Huddinge, Sweden). 5¢-RACE and reverse transcription of total RNA were performed with a kit (5¢-RACE system for rapid amplification of cDNA ends) from Life Technologies, who also provided RNA (0.24–9.5-kb) and DNA ladders (1-kb). Cycle sequencing kits were: Thermo Sequenase for radiolabeled ddNTPs from Amersham Pharmacia Biotec; and ABI Prism Big-Dye terminator from PerkinElmer. Equipment for protein purification was as described previously [19]. Endoglycosidase F/N-glycosidase F and O-glycosidase were from Boehringer-Mannheim. Polyvinyl- difluoride membranes (ProBlott) were from Applied Bio- systems. Purification Mn-LOX was isolated from Ggt and Gga, and purified by chromatography as described before [19,21]. We purified the enzyme from two sources, as the genomic library was obtained from Gga and internal peptides were from Mn-LOX of Ggt. Enzymatic deglycosylation was performed as described previously [19]. Total amino-acid composition The peak fraction of Mn-LOX-Ggt from the gel filtration column was analyzed directly for total amino acids [21], whereas an additional step was used for Mn-LOX-Gga. After gel filtration, this enzyme was purified by SDS/PAGE and blotted onto polyvinyldifluoride membranes. Elec- trophoretic transfer (Mini Trans-Blot, Bio-Rad) was in 10 m M 3-[cyclohexylamine]-1-propane sulfonic acid (pH 11) with 10% methanol (v/v) (100 V, 4 h at 21 °C). The membranes were stained for proteins with Coomassie blue [29]. The excised protein band of Mn-LOX was subject to total amino-acid analysis. Amino-acid sequencing Purified Mn-LOX from Ggt was subject to in situ digestion in the SDS/PAGE gel with Lys-C, trypsin, and with V8 protease [30]. Peptides were isolated by narrow-bore RP- HPLC on the Smart System (Amersham Pharmacia Bio- tech) and subject to amino-acid sequencing (PerkinElmer ABI 494 Sequencer). Analysis of the C-terminal amino-acid sequence was performed as described previously [31]. RT/PCR analysis and cloning Mycelia of G. graminis were harvested by filtration. Total RNA was prepared by grinding of mycelia in liquid nitrogen, extracting with the RNeasy plant kit, and checking for integrity by agarose gel electrophoresis. About 2.5 lg of total RNA and 1 U enhanced avian reverse transcriptase (Sigma) in 20 lL were used for first-strand synthesis (55 °C for 50 min) according to the manufac- turer’s protocol, and 4 lL were used as templates for each PCR. For 5¢-RACE, total RNA (1 lg) was transcribed with a gene-specific primer (Mns21r, 5¢-CTGGCTGG GGGGTGTACTTCTTCT-3¢) according to the protocol from Life Technologies for 5¢-RACE of GC-rich templates. The PCR (50 lL) contained 0.4 l M each primer, 10 m M Tris/HCl pH 8.3, 50 m M KCl, 3.0 m M MgCl 2 ,0.2m M dNTPs and 1.5 U Taq DNA polymerase. The PCR protocol was: 94 °C for 3 min, 1 cycle, followed by 94 °C for 45 s, 48 °C for 45 s, 72 °Cfor1minfor30 cycles, a final extension step (72 °C, 10 min) and then cooling to 8 °C. The amplicons were cloned into the TA vector pCR2.1-TOPO and used for heat shock transfor- mation of Escherichia coli (TOP10, Invitrogen). Sequencing was performed by the cycle sequencing method. Ó FEBS 2002 Cloning of manganese lipoxygenase (Eur. J. Biochem. 269) 2691 Genomic library screening The genomic library of Gga was constructed by partial digestion of genomic DNA with TspeI and ligated into the EcoR1 site of k-ZAP II (Stratagene) as described previously [27]. A cDNA probe (0.33 kb) was generated by RT/PCR using primers MnS2 and MnS1 and labeled with 32 Pusing the random priming method [32]. Hybridization screening of the genomic library was performed in QuikHyb (Strat- agene) as described [28,32]. Three rounds of screening purified positive plaques. The Bluescript plasmids were rescued from the Bluescript SK phagemid with helper phage (Stratagene). Restriction analysis Analysis of the Bluescript plasmids was performed with restriction enzymes followed by size-fractioning in 0.8–1.5% agarose gels. SpeIandNsiI yielded a DNA segment ( 3 kb), which contained the coding region of the Mn-LOX gene. This segment was subcloned into pGEM- 5Zf(+) (Promega). Northern and Southern blot analyses Total RNA (15 lg) was size-fractionated by electrophoresis in 1% agarose/0.22 M formaldehyde gels, transferred to Hybond-N membranes and hybridized in QuikHyb (Strat- agene) with the 32 P-labeled cDNA probe (337 bp, see below) as described previously [32]. The DNA fragment, which was obtained by cleavage of the genomic sequence of Mn-LOX with BamHI and NotI was used as a probe (see Fig. 1). Genomic DNA of Gga was isolated and  1.7 lg was digested with NotIandHindIII. Homology search The gapped BLAST algorithm of the GenBank at NCBI (http://www.ncbi.nlm.nih.gov; [33]) was used for database search and for pair wise alignments, whereas the LASERGENE MEGALIGN program (Dnastar, Madison, WI, USA) was used for multiple alignments. RESULTS Amino-acid analyses and degenerate oligonucleotides Native Mn-LOX-Gga was purified to homogeneity and had an apparent molecular size of 90–110 kDa on SDS/PAGE, whereas Mn-LOX-Ggt appeared to be larger (100– 140 kDa) [19]. After N- and O-linked deglycosylation, SDS/PAGE of Mn-LOX showed two bands of  67 and  73 kDa. Mn-LOX-Gga yielded mainly the 67 kDa pro- tein, whereas Mn-LOX-Ggt yielded both with equal inten- sity, possibly due to incomplete deglycosylation [34]. The total amino-acid compositions of Mn-LOX-Ggt and Mn-LOX-Gga and of the deduced precursor proteins are summarized in Table 1. The four C-terminal amino acids were determined by C-terminal sequencing as FLSV. In situ digestion of Mn-LOX- Ggt with endoproteinase Lys-C, V8 and trypsin followed by peptide separation and amino-acid sequencing [30] yielded 10 relatively long internal peptide sequences (including the C-terminal peptide of 23 amino acids). Two peptides were successfully used for design of degenerate oligonucleotide primers: peptide-1, LYTPQPGRYAAACQGLFYLDARS NQFLPLAIK (obtained with Lys-C) was used to design the sense primer Mn60 (5¢-AACCAGTTCCTSCCSCTCGCS ATCAA-3¢) and the antisense primer Mn15R (5¢-GTCGA GGTAGAAGAGGCCCTGRCAVGC-3¢), whereas the tryptic peptide-2 (HPVMGVLNR) provided the sense primer EO3a (5¢-CATCCSGTSATGGGYGTSCTBAA-3¢) and the antisense primer EOr3a (5¢-CGGTTSAGGACRC CCATVACVGGRTG-3¢). The internal peptide sequences of the remaining eight peptides (the C-terminal peptide, GLSQGMPFWTALNPAVNPFFLSV; VDDAFAAPDL LAGNGPGRA; EMAGRGFDGGLSQG; TNVGADLT YTPLDD; FSGVLPLHPAWL; QAVEQVSLLAR; GLV GEDSGPR; LFLVDHSYQK) could be identified in the deduced protein sequences of Mn-LOX (Fig. 2). RT-PCR cDNA was initially prepared from Ggt. The primers Mn60 and EOr3A generated a band of 230 bp, which contained Fig. 1. Organization of the Mn-LOX-Gga gene, Northern and Southern blot analyses. (A) The Mn-LOX-Gga gene. The open box indicates the protein coding region. The solid lines show the 5¢-and3¢-UTR and the intron. An arrow marks start of transcription and some restriction sites are marked. A solid line shows the two overlapping cDNA fragments, which were obtained by RT/PCR and used for screening of the genomic library. (B) Northern blot analysis of Ggt yielded a major signal at  2.7 kb and a minor signal at  3.1 kb. Size markers are from the RNA ladder. (C) Southern blot analysis of Ggt.Genomic DNA was digested with BamHI and NotI,whichwereexpectedto yield a 1.4 kb fragment. The latter was detected as shown. Size markers are shown by arrows. 2692 L. Ho ¨ rnsten et al. (Eur. J. Biochem. 269) Ó FEBS 2002 the deduced sequence WLLAK, which is well conserved in LOX [35], in one of the reading frames, whereas the primers EO3A and Mn15R generated a band of 220 bp. Misprim- ing of the EO3A primer formed the latter, as a sense primer from this sequence (MnS2: 5¢-CCGTTCAGCGTCGAGA GCAAGG-3¢) and an antisense primer from the other sequence (MnS1, 5¢-TCTCGGGGATCGTGTGGAAGA GCA-3¢) amplified a fragment of 337 bp. The latter contained WLLAK and the amino-acid sequence of pep- tide 1 in one of the reading frames. This amplicon was used as a probe for screening of a genomic library of Gga and for Northern blot analysis. Isolation of genomic clones About 100 000 plaques were screened with the cDNA probe and 11 positive clones were obtained. Positive plaques were subject to three rounds of plaque purification. All rescued Bluescript SK phagemids seemed to contain the same insert of  8 kb as judged from restriction enzyme analysis. Organization of the Mn-LOX-Gga gene A map of the Mn-LOX-Gga gene is shown in Fig. 1A, and important features are summarized in Table 2. About 3.4 kb of the genome of Gga was sequenced,  0.8 kb of the 5¢-untranslated region (5¢-UTR) (up to the vector sequence) and  0.6-kb of the 3¢-UTR. The GC content averaged 60.5%. The 5¢-UTR did not contain TATA or CAAT-like boxes. The transcription start point for the Mn-LOX-Gga and Mn-LOX-Ggt genes were determined by 5¢-RACE (Table 2) and found to be located 72 nucleo- tides from the tentative translation start point. About 80% of fungal genes have a purine (usually A) at position )3 from the translation start point [36]. The Mn-LOX-Gga gene had A in this position, whereas the Mn-LOX-Ggt gene had G (Table 2). cDNA analysis also showed the presence of an intron of 133 bp. The exon/intron borders followed the gt/ag rule. There was a typical signal (TGCTAAC; consensus c/TNCTA/GAC/t) for branching that occurs in splicing of RNA of filamentous fungi located 25 nucleotides from the 3¢ acceptor. The intron was short, a characteristic of filamentous fungi [36]. Table 1. Total amino-acid compositions of Mn-LOX and their precur- sors. Amino acids Mn-LOX-Ggt a Mn-LOX-Gga a Measured 618 (602) Deduced b 618 (602) Measured c 618 (602) Deduced 618 (602) Ala Arg Asx Cys Glx Gly His Ile Leu Lys Met Phe Pro Ser Thr Trp d Tyr Val 65 (64) 34 (33) 61 (59) 3 43 (42) 57 (55) 14 17 58 (56) 24 (23) 6 31 (30) 42 (41) 46 (47) 48 (47) 10 21 (20) 37 (36) 74 (70) 40 (37) 65 1 45 53 15 (14) 17 (15) 66 (65) 21 11 (10) 33 (32) 38 35 (34) 36 8 23 37 (35) 67 (66) 26 (25) 59 (57) ND 56 (55) 63 (61) 8 21 69 (67) 18 7 33 45 (43) 40 (39) 36 (35) ND 23 (22) 38 (37) 74 (70) 40 (37) 64 1 45 53 15 (14) 20 (18) 67 (66) 22 10 (9) 33 (32) 38 35 (34) 35 8 23 35 (33) a Normalized to 618 and to 602 amino acids, as the mature proteins may consist of 602 amino acids due to cleavage of a signal peptide (MRSRILAIVFAARHVA) [38]. b Deduced Mn-LOX precursor from partial sequencing of cDNA of Ggt and the sequenced C-terminal peptide (Fig. 2). c Analysed after blotting to poly(viny- lidene difluoride) membranes, which may give artificially low values for Arg, His, and Lys [48]. Fig. 2 The predicted amino-acid sequence of the Mn-LOX-Gga precursor. Amino acids are numbered beginning with the methionine residue (Met1) of Mn-LOX-Gga. Internal peptides generated by cleavage of Mn-LOX- Ggt with endoproteinases are underlined. The amino-acid sequence of Mn-LOX-Ggt differed from Mn-LOX-Gga in only seven positions (K52N, V258A, I384V, I473V, L493V, A507T, and I586M). Table 2. Translation, transcription and termination sequences of the Mn-LOX-Gga gene and the exon–intron borders. Transcription start point a Translation start point a Translation end a 1 gcaggttc… acaaaA 73 TGCGC……GAGCGTC 2058 taaagg Met 1 ArgSerArgIle……PheLeuSerVal 618 Intron 5¢-Donor Branch signal 3¢-Acceptor Intron I …AGCg 445 tatgtgc t 562 gctaac ggctatag 577 CGT… …IleThrSer 124 Arg 125 GlyGlyPhe… a The transcription start point of Mn-LOX-Ggt gene was a(1)gtaggttc…, and the translation start was …acgaaA(73)TGCGC. Ó FEBS 2002 Cloning of manganese lipoxygenase (Eur. J. Biochem. 269) 2693 Northern and Southern blot analyses The cDNA probe hybridized to two poorly separated bands of  2.7 and  3.1 kb, respectively, of total RNA from Ggt (Fig. 1B). The polyadenylation sites were not determined, but the sequenced 0.6 kb of the 3¢-UTR of Mn-LOX-Gga contained three tentative eukaryotic polyadenylation signals [37], i.e., A(2174)AUUAA, A(2438)AUAAC, and C(2577)AUAAA. Southern blot analysis yielded the expec- ted signal at  1.4 kb (Fig. 1C), which was in agreement with a single Mn-LOX gene, but this was not investigated further. Deduced amino-acid sequences The predicted amino-acid sequence of the Mn-LOX-Gga precursor based on an open reading frame of 1854 nucleotides is shown in Fig. 2, and it contained the 10 sequenced peptides as shown. The precursor thus contained 618aminoacidsandhadamolecularmassof67.7kDa. The gene was isolated from a library of Gga, whereas peptide information was obtained from Mn-LOX-Ggt. We also partially sequenced cDNA of Ggt from the 5¢-end (598 amino acids). In combination with the sequenced C-terminal peptide of 23 amino acids of Mn-LOX-Ggt (Fig. 2), we obtained the complete amino-acid sequence. The two proteins were almost identical, as Mn-LOX-Ggt andMn-LOX-Ggadifferedonlyatsevenaminoacids,two of which were found in the sequenced peptides: Ile384 [in GLV(384)GEDSGPR] and Ile586 [in EM(586)AGRGFD GGLSQG]. The N-terminal sequence of Mn-LOX sequences was recently reported in a patent [38], which showed that an N-terminal peptide was released by cleavage between Ala16 and Ala17. This tentative signal peptide, MRS- RILAIVFAARHVA, thus contains four alanine and three arginine residues, which might explain the low number of alanine and arginine residues in the two native Mn-LOX compared to the number of these residues of their deduced precursors (Table 1). Sequence homology When the predicted amino-acid sequence of Mn-LOX-Gga was subject to BLAST search [33], the program reported homology with the consensus sequence of lipoxygenases with three-dimensional structure (Pfam 00305, LOX; http://www.ncbi.nlm.nih.gov/Structure/cdd) and the family of lipoxygenases. Mammalian LOX yielded the highest scores, followed by the plant lipoxygenases. A partial alignment of Mn-LOX with the consensus Pfam LOX is shown in Fig. 3. The BLAST algorithm (with Blosom62) aligned Mn-LOX with 372 residues of mouse 8S-LOX with 26% identical and 42% similar amino-acid residues. The corresponding figures for mouse 12S-LOX of leukocyte- type was 27% and 40% (out of 434 residues). The first 125 amino acids of Mn-LOX showed little homology to LOX; only the coral 8S-LOX indicated homology of this region. The coral LOX could be aligned with more than 500 amino acids (residues 83–587) of Mn-LOX with 24% identical and 39% positive residues. LOX2 of Arabidopsis thaliana andotherplantLOXcouldalsobealignedwith about 25–27% amino-acid identity over 300–400 amino acids, and so could the probable LOX of Pseudomonas aeruginosa [39]. The homology of Mn-LOX and the LOX gene family included the two a helices of the latter that contain the four Fe(II) ligands. These ligands are two His residues found in ahelix 9 (and in the characteristic sequence of 30 amino acids: WLLAK-X 15 -H-X 4 -H-X 3 -E), one His residue and a distant ligand (usually an asparagine residue) in ahelix 18 (and in the characteristic sequence WI-X 4 -H-X 3 -N-X-GQ) [11–15,40]. The region of Mn- LOX that correspondeds to a helix 9 of LOX, contained the sequence WLLAK-X 14 -H(290)-X 3 -H(294)-X 3 -E (3rd line in Fig. 3). Although this sequence contained only 28 amino acids, the two critical histidine residues appeared to be conserved, suggesting that His290 and His294 are Mn(II) ligands. The region of Mn-LOX, which corres- ponded to a helix 18, contained the sequence WI-X 4 - H(478)-X 3 -N(482)-X-G (seventh line of Fig. 3), suggesting that His478 and Asn482 may have the same function as these residues have in Fe-LOX. Finally, there appeared to be conserved amino acids at the C-terminal end of Mn-LOX (data not shown), but the characteristic C-terminal isoleucine residue of LOX was not conserved. The C-terminal amino acid of Mn-LOX was valine, and there is precedence for the C-terminal valine as a metal ligand in both native and recombinant Fe-LOX (see below). Fig. 3. Partial alignment of Mn-LOX-Gga precursor with Pfam LOX. The BLAST algo- rithm was used for alignment. The two sequences were aligned from regions corres- ponding to the beginning of a helix 6 to the end of a helix 21 of soybean LOX L1. Red letters mark identity, blue letters similarity, and letters in italics mark low complexity. The Fe ligands of Pfam LOX shown in this align- ment are His341, His346, His533 and Asn537, which were aligned with His294, His297, His478 and Asn482 in the Mn-LOX sequence. 2694 L. Ho ¨ rnsten et al. (Eur. J. Biochem. 269) Ó FEBS 2002 DISCUSSION We have cloned and sequenced the gene of Mn-LOX of G. graminis and the corresponding cDNA. Our main finding is that Mn-LOX belongs to the LOX gene family with a unique difference in one of the conserved regions. The deduced protein sequence of the Mn-LOX precursor contained 618 amino acids (67.7 kDa). The Mn-LOX precursorissmallerthanmammalianLOX( 73 kDa) andmuchsmallerthanplantLOX( 90 kDa). The central part of the deduced protein sequence of Mn-LOX (over 450 amino acids) could be aligned with  27–28% amino-acid identity and  40% similarity with mammalian LOX (e.g. 15-LOX type 2, 8 and 12-LOX), and with the three-dimensional consensus LOX sequence (Pfam LOX; Fig. 3). Plant and mammalian LOX contain a small N-terminal b-barrel domain and a large C-terminal and catalytic domain, as revealed by three-dimensional analysis [11,15]. The function of the b barrel is unknown, but it is identical in connectivity to the C-terminal domain of certain lipases and might be related to lipid binding [15] and to membrane translocation [41]. The b barrels of plant LOX consist of  150 amino acids and the b barrel of mammalian LOX contains  125 amino acids. In spite of the conserved three- dimensional structure, the plant and mammalian amino- acid sequences of the bbarrel domains cannot be aligned with significant homology [15]. It was therefore not unexpected that the N-terminal 85 amino acids of Mn-LOX failed to align with plant and mammalian LOX. Three- dimensional analysis is needed to determine whether Mn-LOX also contains an N-terminal b barrel. Alignment of the C-terminal domain of Mn-LOX along a helices 9 and 18 of soybean LOX L1 suggested that Mn(II) could be coordinated to four amino acids in the same way as Fe(II) in LOX (cf. Fig. 3). The four Mn ligands were tentatively identified as His290 and His294 of a helix 9 and His478 and Asn482 of a helix 18. In LOX L1, this asparagine residue is located with its amide oxygen at 3.05–3.3 A ˚ from Fe, in the various structures, and thus is considered to be only a weak ligand. By further analogy, the fifth Mn ligand could be the carboxyl oxygen of the C-terminal amino acid valine. Rat 5-LOX has valine as the C-terminal amino acid [42], and site-directed mutagenesis of murine platelet and leukocyte 12-LOX has shown that the C-terminal isoleucine may be substituted by valine with retention of enzyme activity, whereas most other substitu- tions yielded inactive enzymes [43,44]. A water molecule is the sixth Fe ligand of soybean LOX-L1 [12], and Fe 3+ OH N has recently been identified as the catalytic base for hydrogen abstraction [45]. This mechanism is also plausible for Mn-LOX, as electron paramagnetic resonance spectra of the X- and W-bands (9.2 and 94 GHz, respectively) show that the coordination environment of Mn-LOX is similar to that in Fe-LOX with three N-ligands to the metal centre and O-ligands in the remainder of the six coordination positions [46]. These data are consistent with nitrogen atoms of the three histidine residues and oxygen atoms of valine, asparagine, and water. X-ray crystallography will be needed to conclusively confirm that His294, His297, His478, Asn482, Val618 and water are ligands of Mn(II), but this seems likely from the well established sequence homology of LOX, the electron paramagnetic resonance spectra, and the precedence of conserved metal ligands in Fe and Mn forms of other homologous enzymes [23–26]. The major part of a helix 9 of soybean LOX L1 with its twoFe(II)ligandscanbealignedwithallknownLOX without amino-acid gaps [13,35,47]. Alignment of a helix 9 with Mn-LOX yielded two single-amino-acid gaps, one between the characteristic motif WLLAK and His290, and the other between His290 and His294 (Fig. 3). As regards a helix 18 and the C-terminal amino acid, there appeared to be no principle differences between Mn-LOX and other LOX. The unprecedented sequence difference in a helix 9 between all published LOX sequences and Mn-LOX is therefore probably of paramount importance for the geometry at the metal center and the metal specificity. It may explain the paradox that Fe- and Mn-LOX can have conserved metal ligands, yet form different oxidation products and abstract hydrogen in different ways [19,20]. It will clearly be of interest to combine molecular modeling with site-directed mutagen- esis of LOX with known three-dimensional structure (e.g. soybean LOX-L1, LOX-L3 or rabbit 15-LOX) to deter- mine the impact of deleting one or two amino acids in a helix 9 on the metal center and on the catalytic properties. The corresponding studies with insertion of one or two amino acids into the a helix 9 of Mn-LOX may also be warranted, but they will only provide circumstantial evidence until the three-dimensional struc- ture of Mn-LOX is solved. Studies on expression of Mn-LOX for this purpose are now in progress. LOX can cause oxidative degradation of cell membranes, and plant LOX are often activated by pathogen attack as a means of pathogen resistance. G. graminis illustrates that an invading pathogen may secrete Mn-LOX as a means of pathogenicity. We report that the unique biochemical properties of Mn-LOX might be related to an unpreceden- ted structural difference in a conserved region near the metal center of Fe-LOX rather than to the metal ligands. ACKNOWLEDGEMENTS Supported by the Swedish Research Council in Medicine (03X-06523) and Magn. Bergvalls Stiftelse. We thank A ˚ . Engstro ¨ m, Uppsala University, for valuable suggestions. REFERENCES 1. Brash, A.R. (1999) Lipoxygenases: occurrence, functions, cata- lysis, and acquisition of substrate. J. Biol. Chem. 274, 23679– 23682. 2. Samuelsson, B. (2000) The discovery of the leukotrienes. Am. J. Respir. Crit. Care. Med. 161,S2–S6. 3. Schaller, F. (2001) Enzymes of the biosynthesis of octadecanoid- derived signalling molecules. J. Exp. Bot. 52, 11–23. 4. Gru ¨ llich,C.,Duvoisin,R.M.,Wiedmann,M.&vanLeyen,K. (2001) Inhibition of 15-lipoxygenase leads to delayed organelle degradation in the reticulocyte. FEBS Lett. 489, 51–54. 5. Go ¨ bel, C., Feussner, I., Schmidt, A., Scheel, D., Sanchez-Serrano, J.,Hamberg,M.&Rosahl,S.(2001)Oxylipinprofilingrevealsthe preferential stimulation of the 9-lipoxygenase pathway in elicitor- treated potato cells. J. Biol. Chem. 276, 6267–6273. 6. Rance ´ , I.I., Fournier, J. & Esquerre ´ -Tugaye ´ , M.T. (1998) The incompatible interaction between Phytophthora parasitica var. nicotianae race 0 and tobacco is suppressed in transgenic plants Ó FEBS 2002 Cloning of manganese lipoxygenase (Eur. J. Biochem. 269) 2695 expressing antisense lipoxygenase sequences. Proc. Natl Acad. Sci. USA 95, 6554–6559. 7. Royo, J., Le ´ on, J., Vancanneyt, G., Albar, J.P., Rosahl, S., Ortego, F., Castan ˜ era, P. & Sa ´ nchez-Serrano, J.J. (1999) Antisense-mediated depletion of a potato lipoxygenase reduces wound induction of proteinase inhibitors and increases weight gain of insect pests. Proc. Natl Acad. Sci. USA 96, 1146–1151. 8. Chan, H.W. (1973) Soya-bean lipoxygenase: an iron-containing dioxygenase. Biochim. Biophys. Acta. 327, 32–35. 9. Wiesner, R., Hausdorf, G., Anton, M. & Rapoport, S. (1983) Lipoxygenase from rabbit reticulocytes: iron content, amino acid composition and C-terminal heterogeneity. Biomed.Biochim.Acta 42, 431–436. 10. Percival, M.D. (1991) Human 5-lipoxygenase contains an essential iron. J. Biol. Chem. 266, 10058–10061. 11. Boyington, J.C., Gaffney, B.J. & Amzel, L.M. (1993) The three- dimensional structure of an arachidonic acid 15-lipoxygenase-I. Science 260, 1482–1486. 12. Minor, W., Steczko, J., Stec, B., Otwinowski, Z., Bolin, J.T., Walter, R. & Axelrod, B. (1996) Crystal structure of soybean lipoxygenase L-1 at 1.4 A ˚ resolution. Biochemistry 35, 10687– 10701. 13. Prigge, S.T., Boyington, J.C., Gaffney, B.J. & Amzel, L.M. (1996) Structure conservation in lipoxygenases: structural analysis of soybean lipoxygenase-1 and modeling of human lipoxygenases. Proteins 24, 275–291. 14. Skrzypczak-Jankun, E., Amzel, L.M., Kroa, B.A. & Funk, M.O. Jr (1997) Structure of soybean lipoxygenase L3 and a comparison with its L1 isoenzyme. Proteins 29, 15–31. 15. Gillmor, S.A., Villasen ˜ or,A.,Fletterick,R.,Sigal,E.&Browner, M.F. (1997) The structure of mammalian 15-lipoxygenase reveals similarity to the lipases and the determinants of substrate specifi- city. Nat. Struct. Biol. 4, 1003–1009. 16. Jisaka, M., Boeglin, W.E., Kim, R.B. & Brash, A.R. (2001) Site- directed mutagenesis studies on a putative fifth iron ligand of mouse 8S-lipoxygenase: retention of catalytic activity on mutation of serine-558 to asparagine, histidine, or alanine. Arch. Biochem. Biophys. 386, 136–142. 17. Hegg, E.L. & Que Jr, L. (1997) The 2-His-1-carboxylate facial triad – an emerging structural motif in mononuclear non-heme iron (II) enzymes. Eur. J. Biochem. 250, 625–629. 18. Herman, R.P. (1998) Oxylipin production and action in fungi and related organisms. In Eicosanoids and Related Compounds in Plants and Animals (Rowley, A.F., Ku ¨ hn, K. & Schwebe, T., eds), pp. 117–132. Portland Press, London, UK. 19. Su, C. & Oliw, E.H. (1998) Manganese lipoxygenase. Purification and characterization. J. Biol. Chem. 273, 13072–13079. 20. Hamberg, M., Su, C. & Oliw, E. (1998) Manganese lipoxygenase. Discovery of a bis-allylic hydroperoxide as product and inter- mediate in a lipoxygenase reaction. J. Biol. Chem. 273, 13080– 13088. 21. Su, C., Sahlin, M. & Oliw, E.H. (2000) Kinetics of manganese lipoxygenase with a catalytic mononuclear redox center. J. Biol. Chem. 275, 18830–18835. 22. Whiting, A.K., Boldt, Y.R., Hendrich, M.P., Wackett, L.P. & Que, L. Jr (1996) Manganese(II)-dependent extradiol-cleaving catechol dioxygenase from Arthrobacter globiformis CM-2. Biochemistry 35, 160–170. 23. Han, S., Eltis, L.D., Timmis, K.N., Muchmore, S.W. & Bolin, J.T. (1995) Crystal structure of the biphenyl-cleaving extradiol dioxy- genase from a PCB-degrading Pseudomonad. Science 270,976– 980. 24. Senda, T., Sugiyama, K., Narita, H., Yamamoto, T., Kimbara, K., Fukuda, M., Sato, M., Yano, K. & Mitsui, Y. (1996) Three-dimensional structures of free form and two substrate complexes of an extradiol ring-cleavage type dioxygenase, the BphC enzyme from Pseudomonas sp. strain KKS102. J. Mol. Biol. 255, 735–752. 25. Boldt, Y.R., Whiting, A.K., Wagner, M.L., Sadowsky, M.J., Que, L. Jr & Wackett, L.P. (1997) Manganese(II) active site mutants of 3,4-dihydroxyphenylacetate 2,3-dioxygenase from Arthrobacter globiformis strain CM-2. Biochemistry 36, 2147–2153. 26. Parker, M.W. & Blake, C.C. (1988) Iron- and manganese-con- taining superoxide dismutases can be distinguished by analysis of their primary structures. FEBS Lett. 229, 377–382. 27. Bowyer, P., Clarke, B.R., Lunness, P., Daniels, M.J. & Osbourn, A.E. (1995) Host range of a plant pathogenic fungus determined by a saponin detoxifying enzyme. Science 267, 371–374. 28. Ho ¨ rnsten,L.,Su,C.,Osbourn,A.E.,Garosi,P.,Hellman,U., Wernstedt, C. & Oliw, E.H. (1999) Cloning of linoleate diol syn- thase reveals homology with prostaglandin H synthases. J. Biol. Chem. 274, 28219–28224. 29. Neumann, U. (1996) Quantitation of proteins separated by elec- trophoresis using Commassie brilliant blue. In The Protein Pro- tocols Handbook (Walker, J.M., ed.), pp. 173–178. Humana Press Inc, Totowa, New Jersey. 30. Hellman, U., Wernstedt, C., Gonez, J. & Heldin, C.H. (1995) Improvement of an ÔIn-GelÕ digestion procedure for the micro- preparation of internal protein fragments for amino acid sequen- cing. Anal. Biochem. 224, 451–455. 31.Bergman,T.,Cederlund,E.&Jo ¨ rnvall, H. (2001) Chemical C-terminal protein sequence analysis: Improved sensititity, length of degradation, proline passage, and combination with Edman degradation. Anal. Biochem. 290, 74–82. 32. Davis, L., Kuehl, M. & Battey, J. (1994) Basic Methods in Mole- cular Biology, 2nd edn. Appleton & Lange, Norwalk, Connecticut. 33. Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Rapp, B.A. & Wheeler, D.L. (2000) GenBank. Nucleic Acids Res. 28, 15–18. 34. Hounsell, E.F. (1998) Characterization of protein glycosylation. Methods Mol. Biol. 76, 1–18. 35. Hughes, R.K., Lawson, D.M., Hornostaj, A.R., Fairhurst, S.A. & Casey, R. (2001) Mutagenesis and modelling of linoleate- binding to pea seed lipoxygenase. Eur. J. Biochem. 268, 1030– 1040. 36. Turner, G. (1993) Gene organization in filamentous fungi. In The Eukaroytic Genome. Organisation and Regulation (Brodo, P., Oliver, S.G. & Sims, P.F.G., eds), pp. 107–125. Cambridge University Press, Cambridge, U.K. 37. Edwalds-Gilbert, G., Veraldi, K.L. & Milcarek, C. (1997) Alter- native poly (A) site selection in complex transcription units: means to an end? Nucleic Acids Res. 25, 2547–2561. 38. Sugio, A., Takagi, S., Christensen, S., Østergaard, L. & Oliw, E. (2002) Lipoxygenase. Patent number WO0220730. 39. Stover, C.K., Pham, X.Q., Erwin, A.L., Mizoguchi, S.D., Warrener, P., Hickey, M.J., Brinkman, F.S., Hufnagle, W.O., Kowalik, D.J., Lagrou, M. et al. (2000) Complete genome sequence of Pseudomonas aeruginosa PA01, an opportunistic pathogen. Nature 406, 959–964. 40. Kuban, R.J., Wiesner, R., Rathman, J., Veldink, G., Nolting, H., Sole ´ ,V.A.&Ku ¨ hn, H. (1998) The iron ligand sphere geometry of mammalian 15-lipoxygenases. Biochem. J. 332, 237–242. 41. Chen, X.S. & Funk, C.D. (2001) The N-terminal Ôbeta-barrelÕ domain of 5-lipoxygenase is essential for nuclear membrane translocation. J. Biol. Chem. 276, 811–818. 42. Balcarek, J.M., Theisen, T.W., Cook, M.N., Varrichio, A., Hwang, S.M., Strohsacker, M.W. & Crooke, S.T. (1988) Isolation and characterization of a cDNA clone encoding rat 5-lipoxy- genase. J. Biol. Chem. 263, 13937–13941. 43. Hammarberg, T., Zhang, Y.Y., Lind, B., Ra ˚ dmark, O. & Samuelsson, B. (1995) Mutations at the C-terminal isoleucine and other potential iron ligands of 5-lipoxygenase. Eur. J. Biochem. 230, 401–407. 2696 L. Ho ¨ rnsten et al. (Eur. J. Biochem. 269) Ó FEBS 2002 44. Chen, X.S., Kurre, U., Jenkins, N.A., Copeland, N.G. & Funk, C.D. (1994) cDNA cloning, expression, mutagenesis of C-terminal isoleucine, genomic structure, and chromosomal localizations of murine 12-lipoxygenases. J. Biol. Chem. 269, 13979– 13987. 45. Tomchick, D.R., Phan, P., Cymborowski, M., Minor, W. & Holman, T.R. (2001) Structural and functional characterization of second-coordination sphere mutants of soybean lipoxygenase-1. Biochemistry 40, 7509–7517. 46. Gaffney, B.J., Su, C. & Oliw, E.H. (2001) Assignment of EPR transitions in a manganese-containing lipoxygenase and prediction of local structure. Appl. Magn. Reson. 21, 413–424. 47. Brash, A.R., Jisaka, M., Boeglin, W.E. & Chang, M.S. (1999) Molecular cloning of a second human 15S-lipoxygenase and its murine homologue, an 8S-lipoxygenase. Their relationship to other mammalian lipoxygenases. Adv. Exp. Med. Biol. 447, 29–36. 48. Ploug, M., Jensen, A.L. & Barkholt, V. (1989) Determination of amino acid compositions and NH 2 -terminal sequences of peptides electroblotted onto PVDF membranes from tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis: application to peptide mapping of human complement component C3. Anal. Biochem. 181, 33–39. Ó FEBS 2002 Cloning of manganese lipoxygenase (Eur. J. Biochem. 269) 2697 . Cloning of the manganese lipoxygenase gene reveals homology with the lipoxygenase gene family Lena Ho¨ rnsten 1 , Chao Su 1 ,. probable LOX of Pseudomonas aeruginosa [39]. The homology of Mn-LOX and the LOX gene family included the two a helices of the latter that contain the four