The Arabidopsis thaliana Isogene NIT4 and Its Orthologs in Tobacco Encode b -Cyano- L -alanine Hydratase/Nitrilase*

,

Among the nitrilases of Arabidopsis thaliana, the first higher plant nitrilases that have been cloned (1)(2)(3), isoform 4 is clearly divergent. The members of the NIT1 group (NIT1, NIT2, and NIT3) are highly similar and share a minimum of 82% identical amino acids. In A. thaliana, they occur clustered on chromosome 3, and although the patterns of expression are distinctly different for each isoform (2)(3)(4)(5), their enzymatic characteristics are very similar, albeit not identical (4). As shown previously (4), a major role for these nitrilases appears to be in the metabolism of nitriles released by breakdown of glucosinolates. A further common feature is that all three isoenzymes can convert indole-3-acetonitrile (IAN) 1 to indole-3-acetic acid (IAA), the plant growth hormone (1)(2)(3). A nitrilase (which iso-form, is at present unknown) may be part of the IAA synthase enzyme complex of A. thaliana, a soluble 160-kDa complex catalyzing the conversion of L-tryptophan to IAA in vitro (6).
Nitrilase 4 is peculiar in that it occurs as a single gene at a different chromosomal location (chromosome 5) (3), does not accept IAN as a substrate (Ref. 7 and this paper), and occurs in plants of different taxonomic position, such as tobacco (8) and rice (GenBank TM accession number AB027054). Nitrilases belonging to the NIT4 family thus may have a different and more general function and will likely not be associated with auxin production.
In a previous study we reported about the enzymatic characterization of the A. thaliana nitrilase subfamily encoded by the NIT2/NIT1/NIT3 gene cluster (4). During this work, we notified that Arabidopsis NIT4 has a quite different substrate specificity compared with the NIT1/NIT2/NIT3 group. Here, we report about the elucidation of the enzymatic function of the NIT4 enzyme family.

EXPERIMENTAL PROCEDURES
Plant Material-A. thaliana ecotype C24 and Nicotiana tabacum W38 were grown in a greenhouse in standard soil at 20°C, 70% relative humidity, and 210 mol photons m Ϫ2 s Ϫ1 for a 16-h photoperiod. Seeds of L. angustifolius were sown on Vermiculite and grown in a growth chamber under the following climatic conditions: 16-h photoperiod, 120 mol photons m Ϫ2 s Ϫ1 , 24°C during photoperiod, 20°C during night, 70% relative humidity.
General Procedures-The following general procedures have been described elsewhere (4): sodium dodecyl sulfate-polyacrylamide gel electrophoresis and protein determination.
Vector Construction and Cloning of Nitrilase 4 cDNAs-All basic molecular techniques were adapted from Ausubel et al. (12) or Sambrook et al. (13). Sequences of polymerase chain reaction-amplified or mutated cDNAs were verified by sequencing. Cloning of NIT4 cDNA was described previously (4). The cDNAs for the N. tabacum nitrilases TNIT4A and TNIT4B were kindly provided by Dr. Kazuo Yamaguchi (Institute for Gene Research, Kanazawa University, Kanazawa, Japan), and cloning into pET-21b(ϩ) (Novagen, Madison, WI) was done as described for A. thaliana NIT4 (4). Mutations were introduced using the GeneEditor in vitro Site-directed Mutagenesis System (Promega, Mannheim, Germany).
Expression and Purification of NIT4 Enzymes-The Escherichia coli strain BL21 (DE3) was used for expression of plant nitrilases. Bacteria grown overnight (600 ml) were collected by centrifugation (5000 ϫ g, 5 min, 4°C) and resuspended in 60 ml of lysis buffer (50 mM sodium phosphate buffer, pH 8.0, 300 mM NaCl, 10 mM imidazole, 5 mM 2-mercaptoethanol, 1 mg ml Ϫ1 lysozyme). Lysis was carried out on ice for 30 min and was completed with six bursts of ultrasound (1 min, 40 watts) using an ultrasound tip (Sonifier B-17, Branson). The 10,000 ϫ g supernatant (10 min, 4°C), containing soluble nitrilase protein, was used for (NH 4 ) 2 SO 4 precipitation (40% saturation), and the precipitate was resuspended in 12 ml of lysis buffer omitting lysozyme. This fraction was used for purification of the hexahistidine-tagged nitrilases using a 0.5-ml column of Ni 2ϩ -nitrilotriacetic acid-agarose (Qiagen, Hilden, Germany). Nitrilase bound to the column was eluted with 50 mM sodium phosphate, pH 8.0, 300 mM NaCl, 250 mM imidazole. One ml of the peak fraction was collected and desalted using a NAP-10 column (Amersham Pharmacia Biotech) which was equilibrated in 50 mM potassium phosphate, pH 8.0, 1 mM DTT. The resulting nitrilase fraction was purified at least to 95% homogeneity as judged by Coomassie Blue-stained SDS gels. Protein concentration varied between 80 and 120 g ml Ϫ1 , and total volume was 1.5 ml. The protein was shock-frozen in liquid nitrogen and stored at Ϫ80°C for a maximum duration of 8 weeks.
Preparation of Plant Extracts-One gram of plant material was ground to a fine powder in a mortar with liquid nitrogen and thawed with continuing grinding in 3 ml of 100 mM potassium phosphate buffer, pH 8.0, 1 mM EDTA, 1 mM DTT. The homogenate was centrifuged (15 min, 10,000 ϫ g, 4°C), and the supernatant was again centrifuged (20 min, 100,000 ϫ g, 4°C). The resulting supernatant (soluble proteins) was brought to 40% saturation of (NH 4 ) 2 SO 4 by adding a 100% saturated (NH 4 ) 2 SO 4 solution dropwise. After stirring on ice for 20 min, precipitated proteins were collected by centrifugation (15 min, 10,000 ϫ g, 4°C) and resuspended in a small volume (0.5-2 ml, depending on the size of the pellet) of 100 mM potassium phosphate buffer, pH 8.0, 1 mM EDTA, 1 mM DTT. Because extracts of blue lupine seedling have a high Asn content, they were desalted in the same buffer using PD-10 columns (Amersham Pharmacia Biotech) before and after the (NH 4 ) 2 SO 4 precipitation.
Colorimetric Determination of Nitrilase and Nitrile Hydratase Activity-Nitrilase activity was determined by analyzing the released ammonia using the Bertholet reaction as described previously (4). In brief, the substrate (3 mM) was incubated with 400 -600 ng of purified protein or 50 -100 g of protein (crude extracts) in 0.05 M potassium phosphate buffer, pH 8.0, at 30°C in a total volume of 1 ml. For background control, heat-denatured enzyme (10 min, 100°C) was used. After the indicated times (10 min to 4 h) aliquots of 0.1 ml were taken, and the reaction was stopped by adding 0.1 ml each of 0.33 M sodium phenolate, 0.02 M sodium hypochlorite, and 0.01% (w/v) sodium pentacyanonitrosyl ferrate(III) (sodium nitroprusside). After heating for 2 min at 95°C, the sample was diluted with 0.6 ml of water, and the absorbance was read at 640 nm. Each experiment was calibrated with NH 4 Cl solutions of known concentrations. For parallel determination of both nitrilase and nitrile hydratase (NHase) activity, two 0.1-ml samples were taken and boiled for 10 min to stop the reaction. Both aliquots were diluted to 1 ml with water, and 0.1 ml of the first sample was used for determination of ammonia (resulting from nitrilase activity). To the second sample, 0.25 units of asparaginase (from Erwinia chrysanthemi, Sigma) was added and incubated for 30 min at 37°C. Subsequently, ammonia was determined using a 0.1-ml aliquot as described above (representing both nitrilase and NHase activity). NHase activity was then calculated from the difference of both samples. The 1:10 dilution was necessary for two reasons. (i) The concentration of nitrilase-and asparaginase-released ammonia was usually outside the linear range of the calibration curve (which was between 0.01 and 0.5 mM). (ii) The asparaginase used showed Ala(CN) hydratase activity when incubated at Ala(CN) concentrations higher 1 mM resulting in overestimated background values because Ala(CN) does not decrease below 1 mM in the background samples. By diluting the sample (and therefore the Ala(CN)) before asparaginase was added, this effect was negligible.
Determination of Nitrilase and Nitrile Hydratase Activity by LC-MS-For some experiments, nitrilase and NHase activity were determined by liquid chromatography coupled to electrospray-ionization mass spectrometry (LC-ESI-MS). After the indicated times, 0.1-ml aliquots were withdrawn from the reaction vessels, and 100% (v/v) ethanol was added to reach a final concentration of 80% (v/v) ethanol. The samples were boiled for 10 min and subsequently centrifuged (15 min at 13,000 rpm in a tabletop centrifuge) to collect insoluble material. The supernatant was evaporated to dryness and subsequently resuspended in 0.5 ml of 0.5 mM pentadecafluorooctanoic acid (PDFOA). After a second centrifugation, the supernatant was transferred to a fresh reaction tube. Reverse phase liquid chromatography of the underivatized amino acids was carried out according to Chaimbault

Nitrilase 4 Isoforms
Are not Restricted to Brassicaceae-Partial or complete cDNA or genomic sequences of nitrilases are known from several plant species like A. thaliana (1-3), Brassica campestris (Chinese cabbage) (15,16), Lotus japonicus (GenBank TM accession number AW720658), N. tabacum (tobacco) (8), and Oryza sativa (rice) (GenBank TM accession number AB027054). By comparing the homologies between these nitrilases ( Fig. 1), they can be divided into two groups. The first group, referred to as NIT1 group, seemed to be specific for nitrilases from Brassicaceae. Because many species of the Brassicaceae are characterized by their high glucosinolate content and glucosinolate-derived nitriles are among the best substrates for NIT1-NIT3, a function of these enzymes in glucosinolate metabolism has been proposed (4). Arabidopsis NIT4 belongs to the second group, further referred to as NIT4 group to which, in addition to the NIT4s of the Brassicaceae, all known nitrilases of other plants belong. This clustering of the NIT4 sequences could also be observed by phylogenetic analysis using the PHYLIP software package (Dr. J. Felsenstein, Department of Genetics, University of Washington, Seattle, WA) (data not shown). The NIT4 homologs therefore have to be considered as orthologs which means that the present NIT4 genes share a common ancestor. It therefore seems likely that members of the NIT4 group may also have a conserved function.
Expression of Enzymatically Active NIT4 of A. thaliana in E. coli-By using exactly the same strategy as described previously for NIT1, NIT2, and NIT3 (4), NIT4 could be expressed in E. coli from its genuine start codon as a fusion protein with a hexahistidine tag joined to the C-terminal amino acid of the enzyme via a Val-Glu dipeptide spacer. Although, as in the case of the other three nitrilases from A. thaliana, most of the bacterially expressed protein was found in the 10,000 ϫ g sediment, a fraction of the recombinant nitrilase could be purified from the soluble protein lysate by metal-chelate affinity chromatography on nickel-nitrilotriacetate columns. The resulting fraction was purified at least to 95% homogeneity as judged by Coomassie Blue-stained SDS gels. We tested more than 25 selected substrates using the purified enzyme because in preliminary experiments we observed striking and qualitatively different results when the purified enzyme or a crude extract of NIT4-expressing E. coli was used. The results show that NIT4 is highly specific for ␤-cyano-L-alanine (Ala(CN)) ( Table I). The activity of NIT4 against 3-phenylpropionitrile (PPN) or allylcyanide, which are the best substrates for NIT1-NIT3 (4), was very low, and indole-3-acetonitrile (IAN), a precursor of the plant hormone indole-3-acetic acid (IAA), was not detectably converted by NIT4. This is in agreement with in planta data of Schmidt et al. (7) and Dohmoto et al. (17) who were unable to elicit an auxin response with IAN in wild type tobacco (which expresses at least two NIT4 homologs) as well as in NIT4-overexpressing tobacco (17). Transgenic tobacco plants expressing either NIT2 (7), NIT1 (17), 2 or NIT3 (17) of A. thaliana converted this substrate to IAA and developed strong phenotypic symptoms of auxin overproduction.
NIT4 Is Both a Nitrilase and a Nitrile Hydratase-Since the enzymatic assay used so far was based on the analysis of released ammonia, it was necessary to show that this ammonia represented the nitrile nitrogen and not the amino nitrogen. Thus, we analyzed the reaction products by thin layer chromatography (data not shown) and liquid chromatography coupled to nanospray-ionization mass spectrometry (LC-ESI-MS) (Fig.  2). The production of aspartic acid (Asp) could unequivocally be shown by co-chromatography with authentic Asp during TLC (data not shown) as well as by LC-ESI-MS ( Fig. 2A) and by its collision-induced decomposition spectrum (Fig. 2B). Unexpectedly, Asn could also be detected and occurred in amounts about 1.5 times higher than Asp. Therefore, NIT4 not only has a nitrilase activity (converting Ala(CN) to Asp) but also a nitrile hydratase (NHase) activity (converting Ala(CN) to Asn). Nitrilases hydrolyze nitriles by the successive addition of two molecules of water, whereas the substrate remains covalently bound to the enzyme's catalytic-site cysteine (18). Our results suggested that Asn may occur as a free intermediate of the nitrilase reaction rather than enzyme-bound. However, this could be ruled out by the observation that Asn is only marginally converted to Asp by NIT4 (Table I). This negligible asparaginase activity of NIT4 (ϳ6 nkat (mg protein) Ϫ1 ) cannot account for the observed levels of Asp that is formed at a rate of ϳ500 nkat (mg protein) Ϫ1 . Furthermore, asparaginase activity was not observed in every preparation and results therefore most likely from small contaminations of the preparations 2 R.-C. Schmidt, unpublished data.  Ϫ1 . Activity not detectable (in alphabetical order): 2-aminobenzonitrile, benzonitrile, 2-chloroacetamide, cyanoacetamide, ␣-cyanocinnamic acid, 1-cyano-1-cyclopropane carboxylic acid, 6-cyanopurine, 2-cyanopyridine, 3-cyanopyridine, 4-cyanopyridine, cyclopropanecarbonitrile, glutamine, 4-hydroxybenzonitrile, 3-hydroxypropionitrile, indole-3-acetonitrile, indole-3-carbonitrile, mandelonitrile, naphthalene-1-carbonitrile, (phenylthio)acetonitrile, and propionitrile.

Substrate
Relative activity with E. coli asparaginase. In addition, time course studies showed that Asp and Asn are formed at the same time with a constant ratio with no detectable turnover of Asn (Fig. 3). If Asn would indeed be an intermediate of the reaction, we would expect it to accumulate before Asp synthesis starts, whereas with the onset of Asp formation its level should decrease. Taken together, these results prove that NIT4 converts Ala(CN) to either Asp or Asn. Asn is no substrate of the enzyme, and thus, Asn is no free intermediate of the nitrilase reaction of this bifunctional enzyme.
Is the Catalytic Site for Nitrilase and NHase Activity the Same?-The known reaction mechanisms of nitrilases and NHases are quite different (for review see Ref. 19). Although the first bind their substrate covalently to a cysteine residue, the latter use a nonheme iron for their activity. The formation of Asn from Ala(CN) by NIT4 may be the result of a "premature" release of Asn during the nitrilase reaction or it may occur at a second active site with NHase activity. The substrate concentration dependence of the two reactions ( Fig. 4 and Table II) revealed that the K m of Ala(CN) for both reactions is very similar if not identical, whereas the maximum velocity (V max ) is higher for the Asn formation. This result suggests that the catalytic center for both reactions might be the same. Both reactions showed the same temperature and pH dependence (Table II), and they were both inhibited by N-ethylmaleimide (100% inhibition at 2 mM). The involvement of a cysteine residue in both reactions was further indicated by their inhibition at higher concentrations of DTT (ϳ50% inhibition at 10 mM). To pinpoint this residue(s), the proposed catalytically active cysteine of NIT4 for the nitrilase reaction (Cys-197) was mutated to alanine. The mutant protein (NIT4C197A) showed no nitrilase activity, and a strongly inhibited NHase activity (ϳ5% of the wild type protein) (Fig. 2) demonstrating (i) the importance of Cys-197 for both activities but also (ii) that the NHase activity does not completely depend on this residue.
Is the NHase Activity of NIT4 Genuine?-In 1995, Dufour et al. (20) showed that a single amino acid substitution (Gln to Glu) in the cysteine protease papain resulted in a novel NHase activity of the mutated protein. The observed NHase activity of A. thaliana NIT4 could therefore be the result of an artificial mutation. Sequence errors could be ruled out because the reported genomic and cDNA sequences of NIT4 from A. thaliana encode the same polypeptide. Additionally, NHase activity was seen with three different NIT4 homologs (see below). A critical factor may be the intentional "mutation" of the C terminus of the enzyme introduced with the His tag (EVHHHHHH) that was used in all tested recombinant NIT4 proteins. To study the influence of the His tag, we expressed NIT4 using its genuine stop codon in E. coli and enriched the protein by (NH 4 ) 2 SO 4 precipitation and gel filtration. NIT4 activity was found in the void volume of a Superdex 200 HiLoad 16/60 column (Amersham Pharmacia Biotech) indicating that the native molecular mass is greater than 600 kDa. The active fractions showed no asparaginase activity under the conditions applied, but both nitrilase and NHase activities could be detected, displaying a ratio of 1:1.25 like the His-tagged protein (data not shown). The NHase activity is therefore an intrinsic property of the NIT4 protein.
␤-Cyanoalanine Hydrolysis Is a Common Feature of NIT4 Proteins-As mentioned above, proteins homologous to NIT4 are also known from tobacco (8) and rice. The cDNAs of the tobacco nitrilases TNIT4A and TNIT4B were kindly provided to us by Dr. Kazuo Yamaguchi (Institute for Gene Research,  Kanazawa University, Kanazawa, Japan). By using the same cloning strategy as used for A. thaliana NIT4, both cDNAs could be expressed in E. coli, and the proteins were purified using Ni 2ϩ -chelate affinity chromatography (data not shown). Both enzymes hydrolyzed Ala(CN) to Asp and Asn (Fig. 2) showing, however, a lower ratio of NHase/nitrilase activity (ϳ1) (Table III) compared with the Arabidopsis enzyme, thus producing relatively more Asp. Interestingly, PPN was a better substrate for the tobacco enzymes than for Arabidopsis NIT4, the ratio of Ala(CN) to PPN consumption was ϳ5-6 times higher with the tobacco enzymes (Table III). As already mentioned, A. thaliana has a high content of glucosinolates, and glucosinolate-derived nitriles are substrates for the Arabidopsis isoenzymes NIT1, NIT2, and NIT3. It is therefore possible that Arabidopsis NIT4 lost PPN hydrolase activity during evolution to avoid uncontrolled hydrolysis of glucosinolate-derived nitriles. ␤-Cyanoalanine Hydrolysis in Plant Extracts-In 1972, Castric et al. (10) reported the characterization of a ␤-cyanoalanine hydratase enriched from blue lupine seedlings (L. angustifolius), which formed Asn from Ala(CN). It is not entirely clear from this paper if the authors also looked for Asp production from Ala(CN) by their enzyme. We prepared extracts from leaves of A. thaliana and tobacco and from seedlings of blue lupine. Enzymatic activity was enriched from the crude extracts by (NH 4 ) 2 SO 4 precipitation (40% saturation). All extracts showed Ala(CN) NHase as well as nitrilase activity (Fig.  5). Under the conditions tested, no asparaginase activity could be detected, indicating that the observed Asp is formed by a nitrilase reaction. The tobacco extract produced Asp and Asn from Ala(CN) in similar amounts, whereas the extract from blue lupine showed an ϳ2 times higher NHase activity. This activity (100 pkat (mg protein) Ϫ1 ) was comparable to the activity described earlier (140 pkat (mg protein) Ϫ1 of a 0 -30% (saturation) (NH 4 ) 2 SO 4 extract) (10). We also detected Ala(CN) hydratase/nitrilase activities ranging from 50 to 400 pkat (mg protein) Ϫ1 in extracts of Bryonia dioica (Cucurbitaceae), Lactuca sativa (Asteraceae), and Lycopersicon esculentum (Solanaceae), displaying NHase/nitrilase ratios from 1 to 1.8 (data not shown).
In a recent report, the activation of the NIT4 gene of A. thaliana during leaf senescence was described (21). We therefore compared Ala(CN) hydrolysis in extracts from leaves of plants beginning to shoot and from leaves of flowering plants showing signs of senescence (Fig. 5). Both enzymatic activities were severalfold higher in extracts from senescent leaves. Interestingly, the ratios of the two activities (NHase/nitrilase) decreased from 3.3 in nonsenescent leaves to 1.2 in senescent leaves. DISCUSSION The occurrence of nitrilases in higher plants is known since 1958 (22), but until now the main interest was directed to their ability to convert indole-3-acetonitrile to the plant hormone indole-3-acetic acid (1,3,22). Genomic or cDNA sequences of nitrilases are now known from A. thaliana, two Brassica species (B. campestris and Brassica oleracea) (15,16), tobacco (N. tabacum) (17), and rice (O. sativa, GenBank TM accession number AB027054). In addition several expressed sequence tag clones exist, which encode nitrilase-like proteins. The nitrilases of A. thaliana were characterized in most detail. A. thaliana possesses four different nitrilase genes that are differentially expressed (2)(3)(4). While nitrilases homologous to Arabidopsis NIT1, NIT2, and NIT3 are only known from Brassicaceae, NIT4 homologs are also found in tobacco, rice, and lotus ( Fig.  1). Phylogenetic analysis indicates that the NIT4 enzymes are orthologs (data not shown), and we would therefore expect them to catalyze the same reaction.
Arabidopsis NIT4 has a high substrate specificity for ␤-cyano-L-alanine (Ala(CN)) ( Table I), which is also effectively converted by the NIT4 homologs of tobacco ( Fig. 2) but is no substrate for Arabidopsis NIT1-NIT3 (4). In contrast to the report by Bartel and Fink (3), we observed no hydrolysis of IAN to IAA by Arabidopsis NIT4. Because we used a relatively low amount of purified enzyme (a maximum of 2400 ng), the minimal activity observable under the conditions applied is about 150 pkat (mg protein) Ϫ1 , which may be not sensitive enough to detect very weak IAN hydrolysis. The detection limit, however, represents 0.02% of the activity measured with Ala(CN) as substrate.
The expected product of nitrilase-catalyzed hydrolysis of Ala(CN) would be aspartic acid (Asp). However, we found that NIT4 not only produced Asp but also, to an even higher amount, Asn (Fig. 2). The formation of Asn from Ala(CN) would formally be an NHase reaction, and Asn may be further hydrolyzed to Asp by an asparaginase activity. Does NIT4 possess an asparaginase activity and is Asn therefore a free intermediate of NIT4-catalyzed Asp formation from Ala(CN)? Although we detected a minor asparaginase activity in some of our NIT4 preparations (Table I) this, in all probability, resulted from minor contaminations of these preparations with E. coli asparaginase (see above). Additionally, this activity is so low that it could never account for the formation of the amounts of Asp observed. Importantly, time course studies neither showed a lag phase in Asp formation nor a turnover of Asn (Fig. 3)   ing from the known mechanism of "classical" NHases. Elucidating this mechanism will be of value for the design of new nitrile-degrading enzymes that could be of use for environmental and industrial applications. A possible explanation for this NHase activity may be a "premature" release of the enzymebound substrate after the addition of the first molecule of water. If the second molecule of water is not delivered fast enough, the amide may be released, whereas, if the water is present, the substrate will be processed further to the acid.
Interestingly, an enzyme showing the same enzymatic characteristics as NIT4 was purified from Pseudomonas sp. 13 in 1983 (23), but no sequence of this protein was reported until now. The reported biochemical data of this enzyme are very similar to NIT4, but its K m for Ala(CN) is about 1 order of magnitude higher compared with Arabidopsis NIT4.
Ala(CN) is a product of cyanide detoxification of plants. It is produced from cyanide and cysteine by cyanoalanine synthase (9). Recent results indicate that cyanoalanine synthase is an enzyme homologous or identical to mitochondrial cysteine synthase (24,25). In most species analyzed, Ala(CN) is then converted to Asn, although in some species it is converted to the dipeptide ␥-glutamyl-cyanoalanine (␥-Glu-Ala(CN)). The enzyme(s) catalyzing the formation of Asn from Ala(CN) (cyanoalanine hydratase ϭ Ala(CN) NHase) were biochemically studied from lupine seedlings in the laboratory of E. Conn in the early 70s (10) and later by Galoyan et al. (11), but until now, genes encoding such enzymes have not been cloned, and characterizations of corresponding enzymes from other plant species are lacking. Is Conn's cyanoalanine hydratase and NIT4 the same enzyme? In several feeding experiments using H 14 CN, radioactive label could be detected in Asn but Asp was not significantly labeled (10,26) arguing against a NIT4-catalyzed reaction in which similar amounts of Asp and Asn would be expected. We therefore tested the ability of plant extracts from A. thaliana, tobacco, tomato, L. sativa, B. dioica, and blue lupine to hydrolyze Ala(CN) and could unequivocally detect Ala(CN) NHase as well as Ala(CN) nitrilase activity in all tested extracts. In extracts of blue lupine the Ala(CN) NHase activity was dominant; nevertheless, nitrilase activity was clearly detectable. It is therefore likely that the Ala(CN) NHase from blue lupine is a NIT4 homolog. This topic is currently under investigation in our laboratory. Our in vitro data are, however, in contrast to the data obtained in vivo from the H 14 CN labeling experiments mentioned above. A possible explanation could be that the turnover rate of Asp is higher than that of Asn in vivo, as observed in cotton roots (26). In this case Asp would not accumulate.
One possible source of cyanide in higher plants is the biosynthesis of the plant hormone ethylene from 1-aminocyclopropane-1-carboxylic acid (27). During this reaction cyanoformic acid is produced which then spontaneously degrades to carbon dioxide and cyanide. Interestingly, the Arabidopsis NIT4 promoter was found to be activated during leaf senescence as shown by Northern blot analysis (20). We detected severalfold higher NIT4 activity in extracts from senescent leaves of A. thaliana compared with extracts from nonsenescent leaves, and the ratio of NHase to nitrilase activity decreased (Fig. 5).
This change may indicate that different enzymes are involved in Ala(CN) metabolism at different developmental stages, but it may also indicate that NHase and nitrilase activity of NIT4 could be regulated independently, e.g. by posttranslational modifications. Whether NIT4 activity or NIT4 expression is connected to ethylene biosynthesis or cyanide production during leaf senescence will be addressed in future studies.
The results presented in this paper clearly show that NIT4 enzymes from A. thaliana and N. tabacum are Ala(CN) hydratases/nitrilases. NIT4 orthologs are known from several different species of quite different taxonomical position, and it is likely that NIT4 homologs may be present in all higher plants. NIT4 is proposed to take part in cyanide detoxification in vivo in cooperation with Ala(CN) synthase. We propose to reserve the gene name NIT4 for Ala(CN) hydratases/nitrilases. Nitrilases from other plants belonging to the NIT4 family should therefore also be called NIT4 independent of the total number of NIT genes present, as previously done for B. campestris (15), tobacco (8), and rice.