Allene oxide synthase (AOS) is the first enzyme in the lipoxygenase (LOX) pathway which leads to formation of jasmonic acid (JA). Two full-length cDNAs of AOS designated as AOS1 and AOS2, respectively, were isolated from barley (H. vulgare cv. Salome) leaves, which represent the first AOS clones from a monocotyledonous species. For AOS1, the open reading frame encompasses 1461 bp encoding a polypeptide of 487 amino acids with calculated molecular mass of 53.4 kDa and an isoelectric point of 9.3, whereas the corresponding data of AOS2 are 1443 bp, 480 amino acids, 52.7 kDa and 7.9. Southern blot analysis revealed at least two genes. Despite the lack of a putative chloroplast signal peptide in both sequences, the protein co-purified with chloroplasts and was localized within chloroplasts by immunocytochemical analysis. The barley AOSs, expressed in bacteria as active enzymes, catalyze the dehydration of LOX-derived 9- as well as 13-hydroperoxides of polyenoic fatty acids to the unstable allene oxides. In leaves, AOS mRNA accumulated upon treatment with jasmonates, octadecanoids and metabolizable carbohydrates, but not upon floating on abscisic acid, NaCl, Na-salicylate or infection with powdery mildew. In developing seedlings, AOS mRNA strongly accumulated in the scutellar nodule, but less in the leaf base. Both tissues exhibited elevated JA levels. In situ hybridizations revealed the preferential occurrence of AOS mRNA in parenchymatic cells surrounding the vascular bundles of the scutellar nodule and in the young convoluted leaves as well as within the first internode. The properties of both barley AOSs, their up-regulation of their mRNAs and their tissue specific expression suggest a role during seedling development and jasmonate biosynthesis.
Therefore, jasmonate signalling has attracted increasing attention which led to the cloning and characterization of cDNAs coding for enzymes of jasmonate biosynthesis. According to the biosynthetic route proposed by Vick & Zimmerman (1984), jasmonates originate from α-linolenic acid via a linoleneate 13-lipoxygenase-(13-LOX)-catalyzed oxygenation which forms (13S,9Z,11E,15Z)-13-hydroperoxy-9,11,15-octadecatrienoic acid (13-HPOT). 13-HPOT is the substrate of several enzymes (see Blée 1998 and references therein), such as hydroperoxide lyase (HPL), reductase, divinyl ether synthase and peroxygenase. The dehydration of the 13-HPOT to an unstable allene oxide by allene oxide synthase (AOS) (E.C.22.214.171.124, formerly hydroperoxide dehydratase) represents the first specific reaction in the so-called LOX pathway leading to jasmonate synthesis. The allene oxide is either chemically hydrolyzed to the dominating α-ketol (about 80%), γ-ketols (about 10%), and racemic cis-12-oxo-phytodienoic acid (OPDA) (about 10%) ( Fig. 1), or enzymatically converted by allene oxide cyclase (AOC) to enantiomeric pure cis-(+)-OPDA ( Hamberg & Fahlstadius 1990), the immediate precursor of JA. The final steps of JA biosynthesis consist of a reduction of OPDA followed by three β-oxidation steps. Whereas nothing is known about localization or the enzymatic nature of the β-oxidation steps, two OPDA reductases, OPR 1 and OPR 2, were purified ( Schaller & Weiler 1997a; Schaller et al. 1998 ). Two genes, OPR1 and OPR2, with 90% similarity were identified on a genomic fragment from Arabidopsis thaliana ( Biesgen & Weiler 1999). OPR 1 shows a significant similarity to the sequence and properties of O. WARBURG’s yellow enzyme of yeast which reduces the double bound in α,β-unsaturated ketones ( Schaller & Weiler 1997b). This, together with the almost exclusive activity towards those OPDA-isomers which are not formed in plants, suggests a less specific function of OPR1 for JA biosynthesis. In contrast, OPR2 exhibits a high specificity towards the (9S,13S)-enantiomer of OPDA ( Schaller et al. 1998 ). Furthermore, the AOC reaction is highly specific for JA biosynthesis in that it produces the cyclic core structure of the jasmonates with the correct stereochemistry of the side chains ( Hamberg & Fahlstadius 1990). This enzyme only converts the AOS product formed from 13-HPOT, which further supports its specificity for the biosynthesis of jasmonates ( Ziegler et al. 1999 ). AOC was purified recently to homogeneity from corn ( Ziegler et al. 1997 ), but physiological data on the importance of that enzyme for the regulation of JA levels are not yet available.
Expression of AOS was shown to be tightly linked with elevated JA content during the wound response in A. thaliana ( Laudert & Weiler 1998), suggesting an important regulatory role for AOS in this process. Further- more, the AOS promoter of A. thaliana was found to be activated preferentially in the abscission zones of floral organs ( Kubigsteltig et al. 1999 ). If one assumes that such AOS expression attributes to elevated JA levels, a role of jasmonates in floral organ abscission may occur as suggested recently ( Kubigsteltig et al. 1999 ). Taken together, the present data on AOS point to a decisive role of this enzyme in JA-dependent stress responses and developmental processes.
Here, we describe two AOSs from barley which are regulated in seedling development. These AOSs are (i) located in chloroplasts; (ii) preferentially expressed in parenchymatic cells of the scutellar nodule and of the leaf base; and (iii) their appearance is correlated with elevated levels of jasmonates. The isolated and characterized AOS cDNAs were also found to be up-regulated by treatments with jasmonates, octadecanoids, glucose or sorbitol, and were unaffected by treatments with abscisic acid (ABA), Na-salicylate (SA), NaCl or pathogen infection.
Cloning and characterization of full-length cDNAs and Southern blot analysis
A cDNA library prepared from barley leaves treated for 24 h and 48 h with JAME was screened with the flax AOS cDNA. Among the isolated and sequenced clones, two full-length cDNAs revealed sequence homology to the flax AOS. Both clones exhibited stop codons before the open reading frames. These clones (designated as pHvAOS1 and pHvAOS2) contained an open reading frame of 1461 bp and 1443 bp, respectively, which encode polypeptides of 487 and 480, respectively, amino acids with a calculated molecular mass of 53.4 kDa and 52.7 kDa, respectively, and an isoelectric point of 9.3 and 7.9, respectively ( Fig. 2). Comparison with published sequences revealed for the HvAOS1 identity to the flax AOS of 53.9% ( Song et al. 1993 ), to the guayule AOS of 52.8% ( Pan et al. 1995 ) and to the Arabidopsis AOS of 53.5% ( Kubigsteltig et al. 1999 ) on the protein level, whereas the identity to the barley AOS2 is 89.6%. The corresponding data for HvAOS2 were 54.1%, 52.5% and 53.7%. Similar to the guayule AOS, both barley AOSs lack a putative chloroplast peptide, thus contrasting from both the flax AOS as well as the Arabidopsis AOS ( Fig. 2).
Whereas the consensus sequence P-V-NKQCAG of the heme binding domain of the CYP74A enzymes, being a variant of the FxxGxxxCxG motif known from most cytochrome P450 enzymes, is largely preserved among all known AOS, the barley AOS1 and AOS2 exhibit an A → P transition. Furthermore, both AOS exhibit a T → S transition in the heme binding domain compared to the Arabidopsis and guayule AOS. Also in the highly conserved motif –G-KIL ( Chapple 1998), both barley AOSs differ from the three other known AOS sequences by a I → V transition ( Fig. 2).
Southern blot analysis of barley DNA performed with a labeled full-length cDNA probe of AOS1 revealed two bands if restricted with HindIII, EcoRV or DraI. Three bands could be detected if the DNA was restricted with XbaI or BamHI ( Fig. 3), which might be due to internal restriction sites for both enzymes. This banding pattern suggests the existence of two genes in the barley genome coding for AOSs.
Overexpression of AOS and identification of the reaction products
The whole coding regions of pHvAOS1 and pHvAOS2 were cloned in the bacterial expression vector pQE31 and transformed into the host cells SG13009. After induction with IPTG, an additional and prominent band of the expected size of 55 kDa (calculated size 53.4 kDa) could be detected in the total protein patterns of both bacterial extracts upon SDS-PAGE, when compared to the protein pattern of non-induced host cells shown for AOS1 in Fig. 4(a). Immunoblot analysis with a rabbit polyclonal antibody raised against the recombinant protein of pHvAOS1 revealed one band at a similar kDa range ( Fig. 4b).
To detect AOS activity, enzyme assays were performed with [1-14C]-13-HPOT as substrate. When the affinity-purified recombinant AOS1 and AOS2 from barley were used in this assay, the formation of γ- and α-ketol and OPDA could be detected either by GC–MS analysis (data not shown) or by radio HPLC analysis ( Fig. 4d,e). These products are known to derive from the hydrolysis of the unstable allene oxide, thus indicating that the recombinant enzymes are able to perform the AOS reaction. Within Fig. 4(c) a control incubation with 13-HPOT is shown.
Enzymatic properties of the recombinant AOS1
The substrate specificity of the recombinant AOS1 from barley was determined for six fatty acid hydroperoxides. For substrates we used the 9- and 13-hydroperoxides of linoleic as well as linolenic acid, and hydroperoxides of two C20 fatty acids, both containing the hydroperoxy group in position n-6, but one derived from arachidonic acid and one containing one more double bond at n-3. As shown in Table 1, all substrates were converted by AOS with rather different kinetics. The lowest affinity was observed with 13-HPOD followed by the 9-hydroperoxides. Particularly remarkable was the almost fivefold increased affinity toward the 13-hydroperoxides from linolenic acid when compared to the linoleic acid derivatives. Furthermore, conversion of the 9-hydroperoxides of linolenic acid and linoleic acid is surprising, and was added to our knowledge of this type of AOS activity for the first time. It will be interesting to test the substrate specificity occurring in vivo by analyzing the endogenous occurrence of the corresponding ketols. The hydroperoxides of the C20 fatty acids had the highest affinity toward AOS, especially when the number of double bonds is increased. With the exception of 9-HPOD, which showed the highest rate of conversion, the reaction velocity decreased as the affinity increased.
Table 1. . Substrate specificity of barley AOS1
Vmax nkat 2 μl−1
The mean values ± SD of at least four replicates are shown.
Barley leaves respond to treatment with sorbitol or metabolizable carbohydrates with the endogenous accumulation of JA ( Lehmann et al. 1995 ), JA amino acid conjugates ( Kramell et al. 1995 ) and octadecanoids ( Kramell et al. 2000 ). This accumulation might be accompanied with an up-regulation of AOS mRNAs. Kinetic analysis of AOS mRNA accumulation was performed with the full-length cDNA coding for AOS1. As expected with a high degree of sequence similarity, the use of the full-length cDNA coding for AOS2 led to identical results in this type of Northern analysis (data not shown). Upon continuous treatments, mRNA accumulations revealed that there was no detectable constitutive expression in water-treated controls. Desiccated leaves and leaves floated on 1 m sorbitol showed a weak, but significant up-regulation, whereas 0.5 m glucose strongly induced AOS mRNA ( Fig. 5a). Exogenous application of JAME, JA, the conjugate of JA with l-isoleucine (JA-L-Ile), OPDA or OPDAME led to an earlier and transient rise of AOS mRNA accumulation which peaked about 4 h after onset of treatment ( Fig. 5a). However, no AOS mRNA accumulation occurred upon treatment with 90 μm ABA, 50 μm SA ( Fig. 5a), 1 m NaCl, 2 m m aspirin, 1 m 6-deoxyglucose or infection of leaves with powdery mildew (data not shown). The AOS mRNA accumulation was followed by a delayed but constant accumulation of the AOS protein up to 72 h of JAME treatment. The same accumulation, but with a further delay and less abundantly, was observed during sorbitol treatment ( Fig. 5b).
Intracellular localization of AOS protein
AOS proteins localized intracellularly thus far were found within chloroplasts, and the corresponding cDNAs were shown to code for a putative chloroplast transit peptide ( Harms et al. 1995 ; Laudert et al. 1996 ; Song et al. 1993 ). Due to the apparent lack of a putative chloroplast transit peptide within both barley AOSs ( Fig. 2), we were interested in showing their intracellular location. Due to the sequence similarity of AOS1 and AOS2, the antibody raised against the recombinant AOS1 recognized both AOSs equally in each immunological technique. Taking advantage of its up-regulation upon JAME treatment, immunoblot analysis was performed for total protein extracts of protoplasts and chloroplasts isolated from barley leaves treated or non-treated with JAME. The AOS proteins occurred at elevated levels in leaf segments treated with JAME for 24 h ( Fig. 6a, LS). After purification of chloroplasts from protoplasts isolated from JAME-treated leaf segments, the protein ratio of protoplasts to chloroplasts calculated via equal amounts of chlorophyll was found to be about 7 : 5. Therefore, the weak elevated amount of AOS protein in chloroplasts compared to protoplasts ( Fig. 6a, CP versus PP) reflects a preferential if not exclusive occurrence in the chloroplast. In contrast, the cytoplasmatically located JIP-23 is dramatically reduced after purification of chloroplasts ( Fig. 6a). Upon treatment with a proteinase, the AOS protein was detected in chloroplast containing fractions at a similar amount as before the treatment ( Fig. 6a, CP versus CP*). Due to the fact that JIP-23 is proteinase K resistant, activity of this proteinase was checked by degradation of Rubisco in a crude extract (data not shown). The apparent location of AOS protein in chloroplasts could be confirmed by immunocytochemical analysis ( Fig. 6b–d). There was some cross-reactivity of the anti-AOS antibody with cell walls in water- as well as JAME-treated tissues ( Fig. 6b,c). Due to the cross-reactivity of anti-AOS antibody to an 18 kDa protein in extracts of leaf segments ( Fig. 6a, LS) appearing identically in extracts of water-treated and JAME-treated tissues, it is highly probable that the cell wall fluorescence was caused by this cross-reactivity. However, a significant fluorescence label was found in chloroplasts of JAME-treated leaf tissues ( Fig. 6c), whereas in water-treated leaf segments no significant fluorescence was detectable ( Fig. 6b). Since there was only one band recognized by the anti-AOS antibody in total protein extracts of isolated protoplasts ( Fig. 6a), the fluorescence is indicative for AOS protein suggesting its location in chloroplasts. The autofluorescence of chloroplasts is shown in cross-sections which were performed for immunolabeling without the first antibody ( Fig. 6d).
Organ- and tissue-specific accumulation of AOS mRNA correlates with elevated JA levels
The most abundant protein of barley leaves accumulating upon JA treatment is a 23 kDa protein which is also formed in distinct tissues such as the scutellar nodule during seedling development ( Hause et al. 1996 ). This might be linked to JA biosynthesis. Therefore, and with respect to the unique role of AOS in the formation of JA, we analyzed JA levels and tissue-specific expression of AOS during seedling development. In a 6-day-old seedling, AOS mRNA accumulated weakly in the root tip but abundantly in the tissues around the scutellar nodule ( Fig. 7). Moreover, the AOS mRNA accumulated in the base of the primary leaf, and the signal decreased up to the leaf blade to undetectable levels. This is in agreement with the undetectable level of AOS mRNA in water-floated leaf segments ( Fig. 5a). To inspect the tissue-specific accumulation of AOS-protein and mRNA in more detail, immunolocalization and in situ hybridization were done with different cross-sections of the mesocotyl of a 6-day-old barley seedling ( Fig. 8). AOS protein was clearly detectable in parenchymatic cells of the first internode ( Fig. 8b, showing section III in (a)), whereas the vascular bundles of the coleoptile, the shoot apex and the scutellum were free of label. This result could be confirmed by in situ hybridization. Using different cross-sections below and above the nodal plate of the scutellum ( Fig. 8a), AOS mRNA was detectable in young leaves, in the second shoot, in the first internode as well as in the parenchymatic cells surrounding the vascular bundles of the scutellar nodule. The coleoptile and the scutellum were free of label. This tissue-specific expression suggests a specific role of AOSs in seedling development. Interestingly, tissues with strong AOS mRNA accumulation such as the scutellar nodule and the leaf base ( Figs 7 and 8) exhibited significantly elevated levels of JA. Furthermore, a gradual decrease of AOS mRNA accumulation from the leaf base up to the leaf tip correlated with a decrease in the JA level. These data support the assumption that the AOS described here functions in JA biosynthesis.
Among the numerous LOX-derived compounds, the allene oxide is dominant since it is the first intermediate of the JA-forming branch of the LOX pathway, and AOS expression seems to be subject to complex control ( Laudert & Weiler 1998).
Although AOS belongs to the family of the cytochrome P450 enzymes ( Song & Brash 1991), there is no requirement for oxygen, NADPH and a P450 reductase. The enzyme also exhibits an unusually high turnover number when compared to other P450s ( Chapple 1998). Due to notable differences in the heme-binding domain FxxGxxxCxG of most P450 enzymes, there are two families of the CYP74 type. The CYP74A family contains all AOS, whereas the CYP74B family thus far contains the HPL of pepper ( Matsui et al. 1996 ) as well as of Arabidopsis ( Bate et al. 1998 ). Even among the three AOS sequences known to date, there are differences with respect to the occurrence of a putative chloroplast transit peptide. Whereas AOS from A. thaliana ( Laudert et al. 1996 ) and flax ( Song et al. 1993 ) carry this sequence, it is not present in the AOS of rubber particles of the desert shrub guayule ( Pan et al. 1995 ). The possible regulatory role of AOS in JA biosynthesis as well as the apparent heterogeneity among the known dicotyledoneous AOS prompted us to isolate one or more AOS cDNA(s) from a monocotyledonous plant to inspect the expression and to investigate a possible further diversification between AOS from the two classes of angiosperms.
The first monocotyledonous AOSs are different from other AOSs
Both full-length cDNAs, isolated from JA-treated barley leaves, could be identified as AOS via its sequence identity to known AOS as well as by activity tests upon bacterial overexpression. The barley enzymes exhibit two notable features: (i) as the guayule AOS, both barley AOSs lack properties indicative for a chloroplast leader such as the absence of D, I and E, MA at the start, or S-enrichment ( Von Heijne et al. 1989 ); and (ii) the consensus sequence P-V-NKQCAG, which represents the heme binding domain in CYP74A enzymes, carries an A → P transition ( Fig. 2). This might be of consequence for the enzymatic properties due to the assumed role of isoleucine in oxygen activation ( Van Wachenfeldt & Johnson 1995).
The lack of a putative chloroplast leader sequence in the guayule AOS is not that unexpected due to its association with rubber particles ( Pan et al. 1995 ), but the function of AOS in rubber particles is not clear. However, the absence of a chloroplast target sequence in both barley AOS was somewhat surprising since in leaf tissues, enzymes of JA biosynthesis catalyzing reactions upstream of OPDA were found in chloroplasts, such as LOX ( Bell et al. 1995 ; Feussner et al. 1995 ; Heitz et al. 1997 ), AOS ( Blée & Joyard 1996; Harms et al. 1995 ) or AOC (B. Hause, unpublished results). Therefore, we also expected a chloroplast location for barley AOS. Both methods used in this study, cell fractionation and immunocytochemical analysis, revealed the occurrence of AOS protein in barley mesophyll chloroplasts. This is consistent with (i) biochemical data on the occurrence of AOS activity in envelope membrane fractions from pea ( Blée & Joyard 1996); (ii) the putative chloroplast signal peptide in the flax AOS ( Song et al. 1993 ) and the Arabidopsis AOS ( Laudert et al. 1996 ); and (iii) the chloroplastic location of the flax AOS upon overexpression in transgenic potato ( Harms et al. 1995 ). In contrast, transgenic tobacco plants, expressing a flax AOS in the cytoplasm by lack of its chloroplast signal peptide, showed elevated JA and OPDA levels upon wounding only ( Wang et al. 1999 ). This points, together with the recently observed location of AOC within chloroplasts (B. Hause, unpublished results), to a common location of AOS and AOC in wild-type conditions. It remains to be elucidated whether the barley AOS occurs in the outer envelope membrane which is the location where chloroplast proteins lacking a corresponding signal sequence are imported via a transit peptide-independent sorting route ( Keegstra & Cline 1999).
The enzymatic properties of the recombinant barley AOS1 revealed a decreasing KM in the following order of substrates: 9-HPOD > 13-HPOD > 9-HPOT > 13-HPOT. It remains to be elucidated whether the interesting activity with 9-hydroperoxides as substrates occurs in vivo. Very similar kinetic parameters were found recently for the AOS of the 100 000 g membrane pellet of the corn seed homogenate ( Ziegler et al. 1999 ). This preference of the AOS1 for 13-HPOT and the dominant occurrence of 13-LOX forms catalyzing formation of 13-HPOT within the barley chloroplasts ( Feussner et al. 1995 ; Kohlmann et al. 1999 ), suggest that the AOS characterized here may function in JA biosynthesis. Furthermore, all treatments which do not lead to endogenous rise of jasmonates failed to give AOS mRNA accumulation. Among them are 1 m NaCl, 6-deoxyglucose, aspirin, SA, ABA and pathogen infection. The inability of ABA to cause AOS mRNA accumulation accords with the fact that in barley there is no synergistic effect of ABA and JA ( Ortel et al. 1999 ).
Up-regulation of barley AOS mRNAs support that these AOSs are part of a complex regulation of LOX-derived compounds
Up-regulation of mRNAs of barley AOSs support that at least one of these AOS forms are related to JA biosynthesis. Floating of leaf segments on sorbitol or glucose solution led to AOS mRNA accumulation ( Fig. 5). Both compounds are known to lead to the accumulation of jasmonates ( Lehmann et al. 1995 ), JA amino acid conjugates ( Kramell et al. 1995 ) or octadecanoids ( Kramell et al. 2000 ). Like the Arabidopsis AOS ( Laudert & Weiler 1998) and the flax AOS ( Harms et al. 1998 ), the expression of the barley AOSs is induced by JA and OPDA as well as by their methyl esters. This points to an increased capacity of the octadecanoid pathway, as also shown by elevated JA levels in transgenic potato plants overexpressing flax AOS cDNA ( Harms et al. 1995 ). However, the lack of complete correlation of AOS mRNA accumulation, AOS protein accumulation, AOS activity and JA/OPDA accumulation observed in Arabidopsis leaves upon JA treatment or wounding ( Laudert & Weiler 1998), suggest a more complex regulation than detected so far. Other hydroperoxide converting enzymes in leaves, mainly HPL, might compete with the AOS for the common substrate 13-HPOT and have to be regarded as a further determinant in the regulation of JA biosynthesis. Furthermore, (9Z,12R,13S)-12,13-epoxy-9-octadecenoic acid, which might be formed via the peroxygenase branch of the LOX pathway ( Blée 1998), is a strong inhibitor of the AOC ( Ziegler et al. 1997 ), indicating that JA biosynthesis might also be regulated by compounds formed in other branches of the LOX pathway.
Interestingly, SA, which was suggested to block the wound-induced JA biosynthesis in tomato leaves up-stream of OPDA ( Peña-Cortés et al. 1993 ) and to act down-stream of JA ( Doares et al. 1995 ), increased the accumulation of AOS mRNA, AOS protein and, most surprisingly, OPDA levels in Arabidopsis leaves ( Laudert & Weiler 1998). These data led to the suggestion that SA may inhibit the release of OPDA from the chloroplast. Here, a further difference of a monocotyledonous AOS from that of a dicotyledonous plant is obvious: in barley leaves, SA as well as aspirin did not induce accumulation of AOS mRNA ( Fig. 5). However, a remarkable shift of LOX-derived compounds into products of the reductase was shown for barley leaf segments which were floated on SA compared to those which were floated on JA or sorbitol ( Weichert et al. 1999 ). Furthermore, SA induces 13-LOX ( Hause et al. 1999 ), which is not linked to the JA formation ( Vörös et al. 1998 ). As a consequence, a remarkable flux of LOX-derived metabolites upon SA treatment might be possible which bypasses the AOS reaction and leads to compounds putatively regulating JA biosynthesis. Such a link of SA and JA might explain their independent action in separate signaling pathways which were repeatedly observed ( Niki et al. 1998 ; Thomma et al. 1998 ).
AOSs are expressed tissue-specifically in seedling development and correlate to JA levels
A role of jasmonates is suggested for pollen development in Arabidopsis since both the JA insensitive mutant coi1 ( Feys et al. 1994 ) and the JA-deficient triple mutant fad3-2 fad7-2, fad8 ( McConn & Browse 1996) are male sterile. The recently shown AOS promoter activity of A. thaliana in maturing pollen and bases of petioles and stipules correspond to these data, thus suggesting a role of JA in floral organ abscission ( Kubigsteltig et al. 1999 ). Organ-specific analysis of JA levels will substantiate this suggestion.
During seedling development of soybean ( Creelman & Mullet 1995) and barley ( Hause et al. 1996 ), there are elevated levels of JA preferentially in photosynthetically inactive tissues. Interestingly, the AOS forms described here were found to be expressed preferentially in parenchymatic cells around the vascular bundles of the scutellar nodule and leaf base ( Fig. 8). According to previous data ( Hause et al. 1996 ) and that shown here, the JA levels are also remarkably higher in these tissues.
As revealed by 13C-NMR spectroscopy, the mesocotyl and its scutellar nodule are known to facilitate transport of metabolizable sugars ( Ishida et al. 1996 ) released from the endosperm into the leaf base, the growth zone of a monocotyledonous leaf. This may cause osmotic stress in these tissues according to the higher osmolality detected ( Hause et al. 1996 ). One may envisage the following scenario: carbohydrates degraded in the endosperm lead to osmotic stress in such cells mainly used for their transport into growing tissues. As a consequence, a stress-induced rise of JA may occur, which might be at least partially caused by a transcriptional up-regulation of AOS in these tissues ( Fig. 8). This could explain the preferential accumulation of jasmonates in photosynthetically inactive tissues during seedling development. As suggested by Creelman & Mullet (1997), the down-regulation of genes coding for the photosynthetic apparatus, a well-known effect of JA ( Wasternack & Parthier 1997; Weiler et al. 1998 ), may attribute to protection of oxidative stress in tissues still lacking radical scavengers like chlorophyll. Interestingly, the JA-induced JIP-23 of barley leaves if heterologous overexpressed in tobacco also attributes to the down-regulation of photosynthetic genes ( Görschen et al. 1997 ). Since JIP-23 is expressed during seedling development in the companion cells of the vascular bundle of the scutellar nodule and leaf base ( Hause et al. 1996 ), JIP-23 may function as a developmentally regulated factor which is able to potentiate the down-regulation caused by jasmonate. It will be interesting to analyze whether the tissue-specific occurrence of AOS and the overall elevation of JA levels shown here accords with a tissue-specific rise of JA and down-regulation of photosynthetic genes.
cis-(+)-OPDA was purified from flax seeds and checked for purity by GC–MS. C-18 fatty acids and C20 fatty acids were obtained from NuChek Prep (Elysian, MN, USA) and Sigma (St Louis, MO, USA), respectively.
Plant material and treatment of leaf segments
Seedlings of barley (Hordeum vulgare cv. Salome) were grown under greenhouse conditions with 16 h light (with a minimum intensity of 130 μmol m−2 s−1) at 25°C and 70% relative humidity for 7 days if not otherwise indicated. Primary leaves were harvested and cut into 5 cm segments starting 1 cm below the tip. Subsequently, they were floated on water or in an aqueous solution containing either 45 μm (±) JAME, 45 μm JA, 50 μm JA-L-Ile, 50 μm OPDA, 50 μm OPDA ME, 1 m sorbitol, 0.5 m glucose, 1 m 6-deoxyglucose, 1 m mannitol, 90 μm (±-ABA, 1 m NaCl, 50 μm SA or 2 m m aspirin for the times indicated. Floating was performed in Petri dishes containing 25 ml solution per 6 segments at 25°C under continuous white light (120 μmol m−2 s−1) provided by fluorescent lamps (Narva, Berlin, NC, USA, 250/01). Desiccation stress (DS) was performed to a loss of 30% original fresh weight followed by keeping the leaf segments in a plastic bag up to the indicated times. Inoculations with powdery mildew conidia were performed as described previously ( Hause et al. 1999 ). Plant material was shock frozen in liquid nitrogen and stored at −80°C until use.
Isolation of AOS cDNAs and its expression in Escherichia coli
A cDNA library prepared from barley leaves treated with 45 μm JAME for 24 h and 48 h was used ( Lee et al. 1996 ). 5 × 105 phages of the cDNA library were screened with the flax AOS cDNA ( Song et al. 1993 ). Pre-hybridzation was performed at 50°C in 6 × SSC, 5 × Denhardt’s reagent, 1% SDS, 100 μg ml−1 salmon sperm DNA for 1.5 h, followed by hybridization at 50°C for 20 h with 5 × 107 dpm ml−1 of labeled flax-AOS cDNA in the pre-hybidization solution. Filters were washed at 50°C in 2 × SSC + 0.1% SDS, 1 × SSC + 1 × 0.1% SDS and 0.5 × 0.1% SDS, for 30 min each. Positive plaques were purified and converted in phagemids by in vivo excision.
Sequence analysis of plasmid and phagemid clones was performed with the dideoxynucleotide chain termination method using the Thermo Sequenase Cycle Sequencing Kit (Amersham). Complete sequencing of both strands were achieved by means of primer walking. The cDNA inserts homologous to AOS were PCR amplified and cloned into the HindIII–BamHI digested histidine-tagged expression vector pQE31 (Quiagen, Hilden, Germany). The pQE31, both with and without the cDNA inserts, were transformed into the host strain E. coli SG 13009. Total proteins of isopropyl-β-thiogalactopyranoside (IPTG)-induced or non- induced cultures were isolated by sonification in a buffer containing 100 m m Tris–HCl, pH 8.0, 10% (w/v) glycerol, 300 m m NaCl, and 0.1% (w/v) Tween 20. Subsequently, samples were centrifuged at 25 000 g for 15 min to remove cell debris. The recombinant proteins were purified by affinity chromatography on Ni-NTA-agarose in cracking buffer. First, AOSs were bound in the presence of 10 m m imidazol, followed by subsequent washings with 20 m m, 60 m m and 100 m m imidazol, respectively, to remove slightly bound protein. Finally, AOSs were eluted with 500 m m imidazol. To raise rabbit polyclonal antibodies, recombinant AOS1 was isolated and purified in a buffer containing 8 m urea. Determination of proteins, their electrophoretic separation and transfer onto nitrocellulose were performed as described previously ( Lehmann et al. 1995 ).
Extractions of RNA and Northern blot analysis
Total RNA of frozen tissues was extracted by phenol: chloroform: isoamyl alcohol 90 : 30 : 6 (v/v) using the modifications of Andresen et al. (1992) . Electrophoresis of 10 μg total RNA per lane (if not indicated otherwise) and Northern blot analysis was performed according to Maniatis et al. (1989) . Blots were hybridized at 65°C for 16 h with 32P-labelled fragments of the barley AOS1 cDNA encompassing the full-length cDNA sequence. Gel loading was checked by comparing ethidium bromide-stained rRNA.
Southern blot analysis
For Southern blot analysis, 10 μg of genomic DNA was digested with the restriction enzymes EcoRI,BamHI, HindIII, EcoRV, DraI and XbaI and were separated by agarose gel electrophoresis. After vacuum transfer onto a nylon membrane, hybridization was performed with the 32P-labelled fragment of the full-length AOS1 cDNA according to Maniatis et al. (1989) .
Analysis of reaction products of AOSs
Two μg affinity-purified recombinant barley AOSs diluted in 0.9 ml 100 m m potassium phosphate buffer, pH 7.0, were incubated with 100 nmol of [1–14C]-13-HPOT (86 000 dpm) for 30 min at room temperature. The reaction was stopped by adding 100 μl of glacial acetic acid and extracted twice with diethyl ether. The combined organic phases were evaporated under a stream of nitrogen and lipids were reconstituted in 100 μl of HPLC solvent (acetonitrile: water: acetic acid, 60 : 40 : 0.1 v/v/v). The HPLC analysis was essentially performed as described previously ( Blée & Joyard 1996) with a Jasco HPLC system (Gross Zimmern, Germany).
Proteins from leaves, protoplasts, isolated chloroplasts or bacterial extracts were extracted according to Meyer et al. (1988) , solubilized in SDS-sample buffer, subjected to SDS-PAGE and used in immunoblot analysis. Immunodetection of AOSs was performed by using an affinity-purified antibody raised against the recombinant AOS1 as primary antibody and anti-rabbit-IgG conjugated with alkaline phosphatase (Boehringer, Mannheim) as secondary antibody. Staining of immunodecorated AOSs was undertaken with p-nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolylphosphate (BCIP).
Isolation of intact chloroplasts
Chloroplasts were isolated as described for barley leaf segments by Hause et al. (1994) . Briefly, barley leaf segments, floated on distilled water for 48 h or on 45 μm JAME for 6 h followed by H2O treatment for the final 42 h, were treated with 20 m m citrate buffer, pH 5.5; 0.5 m sorbitol, 1% cellulase Y-C and 0.1% pectolyase Y-23 (Seishin, Japan) at 30°C for 3 h. Protoplasts purified in a discontinuous Percoll gradient were lysed in 25 m m Tricin-HCl, pH 8.0 and 0.5 m sorbitol, and chloroplasts were purified by various centrifugation steps in discontinuous Percoll gradients. Treatment of purified chloroplasts with proteinase was performed by incubation with 0.2 mg ml−1 proteinase K for 30 min at room temperature following Percoll gradient centrifugation. The chloroplast fractions were characterized using the following criteria: (i) amount, by chlorophyll content; (ii) intactness, inspected by loss of ribulose-1,5-bisphosphate carboxylase/oxygenase using gelelectrophoretic separation of proteins; and (iii) contaminations, by immunodetection of JIP-23, a non-plastidic protein of barley leaves ( Hause et al. 1994 ).
In situ hybridization and immunocytochemistry
Leaf segments treated for 48 h with water or 45 μm JAME in the dark and tissue around the scutellar nodule of 6-day-old light grown seedlings (see Fig. 8a) were fixed, embedded in PEG and cut as described previously ( Hause et al. 1996 ).
For in situ hybridization, 5 μm thick cross-sections were mounted on poly l-lysine-coated slides, rinsed in Tris–HCl, pH 8.0, and incubated in 1% bovine serum albumin (BSA) in the same buffer for 1 h. After acetylation, sections were dehydrated by a graded series of ethanol and air-dried. For hybridization, a solution of 0.3 m NaCl, 10 m m Tris–HCl, pH 7.5, 5 m m EDTA, 1 × Denhardt’s solution, 50% formamide, 2 mg ml−1 tRNA, and 200 U ml−1 RNase inhibitor containing denaturated DIG-labeled sense or antisense AOS1-RNA was applied and slides were incubated in a humid box at 45°C overnight. After two washing steps with 0.2 × SSC at 55°C for 30 min each, slides were incubated with 20 μg ml−1 RNase A at 37°C for 30 min, followed by washing with 0.2 × SSC at 55°C for 1 h. Immunological detection of DIG-labeled RNA-hybrids was performed with anti-DIG-fab fragment conjugated with alkaline phosphatase (Boehringer, Mannheim, Germany) according to the supplier’s protocol. For localization of AOS protein within the mesocotyl of barley seedlings, 2 μm thick sections of the same tissues used for in situ hybridization were immunolabeled with the purified rabbit-anti-AOS antibody (diluted 1 : 50 in PBS containing 5% (w/v) (BSA) followed by anti-rabbit-IgG antibody conjugated with alkaline phosphatase as described previously ( Hause et al. 1996 ). Staining procedure was performed with NBT and BCIP. Finally, slides were analyzed by bright field microscopy with a Zeiss ‘Axioskop’ microscope (Zeiss, Germany) equipped with a CCD camera (Sony, Japan).
Cross-sections of leaves (2 μm thickness) were immunolabelled by incubation with purified rabbit anti-AOS antibody (diluted as mentioned above) followed by Tyramide Signal Amplification (NEN) according to the supplier’s instructions. Due to the use of a purified antibody it was not suitable to check the specificity via pre-immune serum. The specificity of the purified anti-AOS antibody is indicated by one band detected in total protein extracts of protoplasts ( Fig. 6a). After immunolabeling, sections were mounted in citifluor/glycerol. Control experiments were performed by omitting the first antibody. The fluorescence of immunolabeled AOSs was visualized with a Zeiss ‘Axioskop’ epifluorescence microscope using the proper filter combination.
Extraction and quantification of jasmonates and octadecanoids
Barley leaf segments and other barley tissues (1 g f.w.) were frozen in liquid nitrogen, homogenized in a mortar and extracted with 5 ml 80% (v/v) methanol. For quantification of JA and JAME, appropriate amounts of (2H6)JA were supplemented to the extract. Ion exchange chromatography on DEAE Sephadex A-25 cartridges, RP-HPLC and GC–MS/SIM was performed according Kr amell et al. (2000) .
Fatty acid hydroperoxides
The n-6 hydroperoxides (13-HPOT, 13-HPOD, 15-HPET and 15-HPEP) were prepared by incubation of the corresponding fatty acids with soybean lipoxygenase (Sigma) ( Hamberg & Gotthammar 1973), whereas the n-10 hydroperoxides (9-HPOD and 9-HPOT) were obtained by incubation of the appropriate precursor acids with tomato fruit homogenates ( Matthew et al. 1977 ).
AOS activity assay
AOS activity was measured at pH 7 (50 m m potassium phosphate) in a total volume of 1 ml. The reaction was initiated by the addition of the fatty acid hydroperoxide and the consumption of the substrates was followed by the decrease in the absorbance at 235 nm, which reflects the disappearance of the conjugated double bond of the hydroperoxides. For the determination of the kinetic parameters of the recombinant barley AOSs, the initial reaction velocity was determined at 14 concentrations in the range of 5–90 μm for each substrate. The KM and Vmax values were calculated directly using the Michaelis Menten equation and by plotting the data according to Eadie–Hofstee. Both methods gave similar results.
We thank Dr O. Miersch for JA measurements; Dr A. Brash for providing us with the flax AOS cDNA; Dr J. Lee for providing us with the cDNA library; B. Ortel, S. Wegener, S. Krüger and M. Krohn for skilful technical assistance; A. Pitzschke for purifying the anti-AOS antibody; and C. Dietel for typing the manuscript. This work was supported by a grant from the Deutsche Forschungsgemeinschaft, SFB 363/C5, to C.W. and H.M., SFB 362/B23 to I.F. and HA 2655/3–1 to B.H.