J Biol Chem, Vol. 275, Issue 13, 9673-9683, March 31, 2000

Molecular Cloning, Genomic Organization, and Biochemical Characterization of Myristoyl-CoA:Protein N-Myristoyltransferase from Arabidopsis thaliana^*

Qungang Qi^§, Raju V. S. Rajala^§¶, William Anderson, Chao Jiang, Kevin Rozwadowski, Gopalan Selvaraj, Rajendra Sharma^¶, and Raju Datla^**

From the  National Research Council of Canada, Plant Biotechnology Institute, Saskatoon S7N 0W9 and the ^¶ Department of Pathology and Saskatoon Cancer Center, University of Saskatchewan, Saskatoon S7N 4H4, Saskatchewan, Canada

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Myristoyl-CoA:protein N-myristoyltransferase (NMT, EC 2.3.1.97) catalyzes the co-translational addition of myristic acid to the amino-terminal glycine residue of a number of important proteins of diverse functions. We have isolated a full-length Arabidopsis thaliana cDNA encoding NMT (AtNMT1), the first described from a higher plant. This AtNMT1 cDNA clone has an open reading frame of 434 amino acids and a predicted molecular mass of 48,706 Da. The primary structure is 50% identical to the mammalian NMTs. Analyses of Southern blots, genomic clones, and database sequences suggested that the A. thaliana genome contains two copies of NMT gene, which are present on different chromosomes and have distinct genomic organizations. The recombinant AtNMT1 expressed in Escherichia coli exhibited a high catalytic efficiency for the peptides derived from putative plant myristoylated proteins AtCDPK6 and Fen kinase. The AtNMT was similar to the mammalian NMTs with respect to a relative specificity for myristoyl CoA amoung the acyl CoA donors and also inhibition by the bovine brain NMT inhibitor NIP₇₁. The AtNMT1 expression profile indicated ubiquity in roots, stem, leaves, flowers, and siliques (1.7 kb transcript and 50 kDa immunoreactive polypeptide) but a greater level in the younger tissue, which are developmentally very active. NMT activity was also evident in all these tissues. Subcellular distribution studies indicated that, in leaf extracts, ~60% of AtNMT activity was associated with the ribosomal fractions, whereas ~30% of the activity was observed in the cytosolic fractions. The NMT is biologically important to plants, as noted from the stunted development when the AtNMT1 was down-regulated in transgenic Arabidopsis under the control of an enhanced CaMV 35S promoter. The results presented in this study provide the first direct molecular evidence for plant protein N-myristoylation and a mechanistic basis for understanding the role of this protein modification in plants.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The enzyme myristoyl-CoA:protein N-myristoyltransferase (NMT)¹ catalyzes the co-translational transfer of myristic acid (a rare 14-carbon saturated fatty acid) from myristoyl-CoA to the amino-terminal glycine residue of a number of cellular, viral, fungal, and oncoproteins with diverse functions (1-3). This co-translational modification has been observed only in eukaryotes (4, 5), and it is required for the full biological activities of several N-myristoylated proteins. For example, N-myristoylation of the  subunit of the signal-transducing guanine nucleotide-binding protein (G protein) increases its affinity for subunits and thereby promotes formation of the heterotrimeric complex (6). The tyrosine kinase pp60^v-SRC requires a myristoyl group to bind to the membrane for its stability and consequent cellular transformation, although the kinase activity of this protein is not myristoylation-dependent (3, 7). It has been shown that N-myristoylation of a variety of structural and nonstructural proteins encoded by several mammalian retroviruses is essential for the viral particle formation (e.g. the Pr55^GAG of human immunodeficiency virus 1) (1-3). Insertional mutagenesis of the yeast NMT1 gene has indicated that N-myristoylation is essential for survival and vegetative cell growth (8).

There is very little information regarding protein N-myristoylation in plants. The evidence that proteins can be labeled with [³H]myristate in the aquatic angiosperm Spirodella oligorrhega (9) and that NMT activity is present in wheat germ (10, 11) suggests that myristoylation occurs in plants. In fact, a number of plant proteins contain putative consensus myristoylation motifs (MGXXX(S/T)-, Refs. 12-24). These include the Pto kinase (12) and the closely related Fen kinase from tomato (13), the ATN1 (14) and APK1 (15) kinases from Arabidopsis, several calcium-dependent kinases (CDPKs) from maize (16), rice (17), tobacco (18), and Arabidopsis (19), calcineurin -like proteins from Arabidopsis (20), and G subunits of the G-proteins from several plant species (21). A myristoylation motif is also found in the predicted amino acid sequences of fatty acid acyl-CoA synthetase gene from Brassica napus (22), a nodule-specific gene family from Alnus glutinosa (23), and the Dem (defective embryo and meristems) gene from tomato (24). These observations imply that protein N-myristoylation may be involved in various plant processes. However, in vivo and in vitro evidence for the myristoylation of these proteins and its potential importance for their functions remain to be established in plants.

NMT genes/proteins have been isolated and well characterized from Saccharomyces cerevisiae (8), Candida albicans (25), mammalian cells (26-28), and Drosophila (29). Multiple sequence alignment of all the known NMT genes revealed a high degree of similarity at both the nucleotide and amino acid levels, and more interestingly, the amino acid residues required for the catalytic activity are highly conserved. Immunofluorescence microscopy studies have shown that NMT is distributed uniformly throughout the cytoplasm of yeast and mammalian cells (30, 31). NMT activities have been reported in both insoluble (3, 28, 32-34) and soluble fractions (35). So far the only report on plant NMT has been from wheat germ in which N-myristoylation was demonstrated in a cell-free translation system (10). The wheat germ enzyme showed a peptide substrate specificity that is broadly similar to yeast NMT, although with a different preference for the amino acid at the 5th position of the cAMP-dependent protein kinase-derived peptide substrate (GNAAAARR, Refs. 10 and 11). However, no detailed biochemical and molecular characterization of plant NMT or its potential importance in plant development or function has been reported. We describe here, for the first time, the cDNA cloning, genomic organization, molecular properties and biochemical characterization of NMT from a higher plant. Furthermore, our antisense transgenic studies with Arabidopsis also show the importance of NMT for normal growth and development.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- gt11 cDNA and EMBL3 genomic libraries of Arabidopsis were obtained from CLONTECH Laboratories Inc. Molecular biology reagents and DNA-modifying enzymes were purchased from Amersham Pharmacia Biotech, Life Technologies, Inc., or Stratagene and used as recommended by the suppliers. The cloning vectors were from Invitrogen (pTrcHis.C) and Stratagene (pBluescript SK(+)). Agarose gel DNA extraction kit was purchased from Roche Molecular Biochemicals. Radioisotope [-³²P]dCTP and [³H]myristic acid (39.3 Ci/mmol) were obtained from NEN Life Science Products. The random priming kit was purchased from Life Technologies, Inc. [1-¹⁴C]Myristoyl-CoA (54.7 mCi/mmol) was obtained from Amersham Pharmacia Biotech. [1-¹⁴C]Decanoic acid (55 mCi/mmol), [1-¹⁴C]lauroyl-CoA (55 mCi/mmol), [1-¹⁴C]palmitic acid (43 mCi/mmol), and [1-¹⁴C]stearic acid (58 mCi/mmol) were purchased from American Radiolabeled Chemicals Inc. The purified bovine brain NMT inhibitor protein NIP₇₁ was prepared as described (36). Unless otherwise indicated, all other chemicals were purchased from Sigma.

Plant Growth Conditions-- Arabidopsis thaliana plants (ecotype Columbia), wild type and transgenic, were grown in controlled growth chambers at a photosynthetic flux of 100-120 µE m²·s¹, 20 °C and a photoperiod of 16 h light/8 h dark.

Isolation of the AtNMT1 cDNA, Genomic Clones, and Their Sequence Analyses-- Unless otherwise indicated, DNA manipulations were performed using standard protocols as described by Sambrook et al. (37). Approximately 500,000 plaques of the Arabidopsis cDNA library were screened initially with a 1.6-kb fragment of human NMT cDNA under moderately stringent hybridization conditions and subsequently with a cDNA fragment (GenBank^TM accession number T21207) from Arabidopsis EST (obtained from ABRC, Ohio) under highly stringent hybridization conditions. The cDNA inserts from plaque-purified phage clones were cloned into pBluescript SK(+) vector. One clone (pSKAtNMT1), which contained the largest cDNA insert, was used for further analyses. Approximately 200,000 plaques from the Arabidopsis genomic library were screened with the AtNMT1 cDNA clone, and two positive clones were identified and designated as AtNMT1-G1 and AtNMT1-G2. The genomic inserts from these clones were subcloned into pBluescript SK(+) vector.

DNA sequencing was done by an Applied Biosystems model 373 DNA Sequencer System with a Taq DyeDeoxy^TM Terminator Cycle Sequencing Kit (Applied Biosystems, Inc.). The nucleotide and deduced amino acid sequences of the full-length cDNA clone, AtNMT1 and the two genomic clones, AtNMT1-G1 and -G2, were compared with sequences available in data banks using the FASTA and BLAST search programs. Sequence alignments were performed with a Clustal PC/Gene Program.

Arabidopsis Genomic DNA Isolation and Southern Blot-- Genomic DNA was prepared from the Arabidopsis leaves according to the method of Dellaporta et al. (38). Five micrograms of genomic DNA were digested using several restriction enzymes, the products separated on a 0.8% agarose gel, transferred to nylon membrane (GeneScreen Plus^R, NEN Life Science Products) using the Vacugene Plus System (Amersham Pharmacia Biotech), and cross-linked using a UV cross-linker (Stratalinker®, Stratagene). Hybridization was carried out using QuickHyb solution according to the supplier's protocol (Stratagene).

Construction of Antisense AtNMT1 Binary Vector and Transformation of A. thaliana-- The full-length of AtNMT1 cDNA was amplified by a pair of primers (5'-CGGGATCCATGGCAGATAACAATTCACC-3' with a BamHI site at the 5' end of the cDNA, and 5'-CATGCCATGGTTATAAGAGAACAAGCC-3' with an NcoI site at its 3' end) using Pfu polymerase (Stratagene). The resulting PCR fragment was digested with BamHI and NcoI and cloned into the respective sites of pBI524 (39) in an antisense orientation, under the expression control of a tandem 35S promoter. After sequence confirmation, the correct antisense cassette was then excised from pBI524 by HindIII and EcoRI and inserted into plant transformation vector pRD400 (40). Transformation of Arabidopsis plants (ecotype Columbia) was performed by a vacuum infiltration method (41). Transformed plants were grown in conditions described as above. Transformants were selected on 50 mg/liter kanamycin and further verified by PCR and Southern blot analysis. The growth and developmental phenotype of the T2 transgenic plants were analyzed.

RNA Isolation and Northern Blot-- Total RNA was isolated from root, stem, leaf, flower, and silique of Arabidopsis using TRIzol reagent (Life Technologies, Inc.). The RNA (10 µg) was fractionated in a denaturing formaldehyde (6.7%), agarose (1.2%) gel electrophoresis and transferred onto nylon membrane (GeneScreen Plus®, NEN Life Science Products). The blots were hybridized with a [-³²P]dCTP-labeled 1.3-kb fragment of AtNMT1 cDNA as a probe using QuickHyb Solution as described by Stratagene's protocol.

Expression and Affinity Purification of Recombinant AtNMT1 and hNMT1 in Escherichia coli-- The DNA fragment corresponding the longest ORF of 434 aa was inserted into an E. coli expression vector (pTrcHisC) to create a derivative recombinant plasmid pTrcHisC-AtNMT1. E. coli DH5 cells with pTrcHisC-AtNMT1 were grown to stationary phase at 37 °C in LB medium containing 50 mg/liter ampicillin and induced by 1 mM isopropyl--D-thiogalactopyranoside for 2 h. The cells were harvested by centrifugation at 10,000 × g for 20 min and suspended in 50 mM Tris-HCl, pH 8.0, containing lysozyme (100 µg/ml), 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), leupeptin (0.5 µg/ml), 0.1% Triton X-100, 10 mM MgCl₂, and 10 µg/ml DNase I. The suspension was incubated for 30 min on ice. The lysate was cleared by centrifugation at 15,000 × g for 20 min, and the supernatant was used for the purification of the expressed protein following the supplier's instructions (Invitrogen). The recombinant human NMT (hNMT1) was prepared and purified from E. coli cells as described (42).

Synthesis of Peptide Substrates for AtNMT-- The following peptides were synthesized based on the amino-terminal sequence of known N-myristoylated mammalian proteins. These were synthesized by the solid phase manual protocol developed by Merrified (43): GNAAAAKKRR (type II catalytic subunit of cAMP-dependent protein kinase), GSSKSKPKR (pp60^SRC), GNASSIKKK (M2 gene segment of reovirus type 3), and GAQFSKTARR (myristoylated alanine-rich protein kinase C substrate). The peptides were purified by CM-cellulose column chromatography and G-25 Sephadex gel filtration. The following peptides based on putative plant N-myristoylated proteins were custom-synthesized by Research Genetics, Inc.: GSKYSKATRR (tomato Fen kinase, Ref. 13), GANHSREDRR (tomato DEM, Ref. 24), GGCFSKKYRR (tobacco NtCDPK1, Ref. 18), and GHRHSKSKKK, AHRHSKSKKK (Gly to Ala substitution) and MGHRHSKSKK (with initiator Met based on the NH₂-terminal sequence of Arabidopsis AtCDPK6, Ref. 19).

Synthesis of AtNMT1 Peptide-- The peptide KNNYVEDDENMFRFNY-CONH₂, corresponding to amino acids 114-129 of AtNMT1, was synthesized (Applied Biosystems model 430 A peptide synthesizer, Alberta Peptide Institute, Canada) utilizing t-Boc N protection and benzyl-type side chain protection. The peptide was coupled to two carrier proteins, bovine serum albumin (peptide to protein ratio 6:1) and keyhole limpet hemocyanin (KLH, peptide to protein ratio 10:1). Two milligrams of KLH-peptide conjugate was dissolved in 1 ml of 8 M urea in a 1.5-ml microcentrifuge tube with occasional vortexing over a period of 4 h and centrifuged for 5 min, and the supernatant was decanted into a 6800 molecular weight cut-off dialysis bag. The KLH conjugate was dialyzed against 0.9% NaCl overnight.

Preparation of BSA-Peptide Affinity Column-- BSA-peptide (10 mg) was dissolved in 10 ml of 0.1 M NaHCO₃, pH 9.6, and 2 g of activated CH-Sepharose 4B (Amersham Pharmacia Biotech) prepared according to the manufacturer's instructions. The resultant slurry was mixed by rotation for 16 h at 4 °C. The Sepharose 4B was pelleted by centrifugation (1000 × g for 5 min), and the extent of BSA-peptide coupling was determined by measuring the protein concentration of the peptide conjugate in the supernatant. More than 90% of the conjugate was coupled to Sepharose 4B. The BSA-peptide-Sepharose 4B was washed with 0.1 M NaHCO₃, pH 9.6, in a Buchner flask, and any remaining reactive groups were blocked by incubating the gel with 0.1 M Tris-HCl, pH 7.4, for 2 h at room temperature. The BSA-peptide-Sepharose 4B was washed alternately with low pH (0.1 M acetate buffer, pH 4.0) and high pH (0.1 M Tris-HCl, pH 8.0) buffers containing 0.5 M NaCl.

Production of Anti-peptide AtNMT Antibody-- Antigen (500 µg) was mixed with 0.5 ml of Freund's complete adjuvant and injected into rabbits (0.25 ml under each shoulder blade and 0.5 ml in one hip muscle (subscapular and gluteal injections)). Two weeks later, 500 µg of antigen was mixed with Freund's incomplete adjuvant and injected. Blood was collected for antibody purification. The BSA-peptide conjugate was used to screen for anti-peptide antibody by an enzyme immunoassay.

The anti-peptide antibodies were affinity-purified by adsorption to a 5-ml BSA-peptide-Sepharose 4B (10 × 1.5 cm) column. After adsorption, the column was washed with 0.1 M Tris-HCl, pH 8.0, containing 0.5 M NaCl. The anti-peptide antibodies were eluted using 100 mM glycine, pH 2.5. The eluted fractions were collected in tubes containing 3.5 ml of 1 M Tris-HCl, pH 8.0, and mixed immediately. The purified anti-peptide antibodies were concentrated by dialysis against 30% polyethylene glycol, suspended in 20 mM Tris-HCl, pH 7.0, 50% glycerol, and 0.02% NaN₃, and then stored at 70 °C at a protein concentration of 0.9 mg/ml.

The specificity of anti-AtNMT1 peptide antibody was examined by immunoblotting reaction. The affinity-purified antibody reacted strongly with the purified E. coli expressed AtNMT1 on immunoblot, whereas the preimmune serum did not recognize the protein (data not shown). This detection was prevented when the affinity-purified anti-peptide antibody was preincubated with the peptide conjugate. Similar results were obtained when either the purified recombinant AtNMT1 or the peptide conjugate was adsorbed onto an enzyme immunoassay plate (data not shown). These results confirm the specificity of the anti-peptide antibody for AtNMT1.

Preparation of Subcellular Fractions-- All operations were performed at 4 °C unless otherwise stated. About 10 g of freshly harvested Arabidopsis tissues from roots, stems, flowers, leaves, and siliques were homogenized in 4 volumes of ice-cold extraction buffer (100 mM Tris-HCl, pH 7.6, containing 300 mM mannitol, 4 mM EDTA, 2 mM EGTA, 1% polyvinylpolypyrrolidone, 10 mM 2-mercaptoethanol, 1 mM benzamidine, 0.2 mM PMSF, and 0.7% µg/ml leupeptin). Subcellular fractionation of the tissue homogenate was accomplished by the modified method of sequential differential centrifugation (44) using SS 34 rotor in a Sorvall RC5C centrifuge (NEN Life Science Products) and a Ti 65 rotor in a Beckman ultracentrifuge (L8-70M Ultracentrifuge, Beckman). The homogenate was filtered through four layers of cheesecloth and one layer of Miracloth. The filtrate was centrifuged at 250 × g for 10 min, and supernatant was re-centrifuged at 500 × g for 20 min to obtain the chloroplast-enriched fraction as the pellet. The supernatant was centrifuged at 4,000 × g for 30 min to obtain the mitochondrial pellet. The supernatant was centrifuged at 20,000 × g for 40 min to obtain the microsomal pellet. The microsomal supernatant was further centrifuged at 100,000 × g for 70 min, yielding the ribosomal pellet and the cytosolic fraction with supernatant. Each of these pellets was washed once using resuspension buffer (100 mM Tris-HCl, pH 7.6, containing 4 mM EDTA, 2 mM EGTA, 10 mM 2-mercaptoethanol, 1 mM benzamidine, 0.2 mM PMSF, and 0.7% µg/ml leupeptin), centrifuged again, and finally redissolved in resuspension buffer. Aliquots from each fraction were used for assaying NMT and organelle marker enzymes, as well as for immunoblotting analyses.

SDS-Polyacrylamide Gel Electrophoresis and Western Blot Analysis-- Proteins isolated from Arabidopsis tissue or purified from E. coli cells expressing AtNMT1 were separated on 10% SDS-polyacrylamide gels according to the procedure described by Sambrook et al. (37). Western blot analyses were performed essentially as described (45). The anti-AtNMT1 antibody and alkaline phosphatase-conjugated secondary antibody (anti-rabbit IgG, Bio-Rad) were diluted to 1:500 and 1:3000, respectively, in TBST (20 mM Tris-HCl, pH 7.5, 1% BSA, 500 mM NaCl, 0.1% Tween 20). The blots were developed using nitro blue tetrazolium in conjunction with the alkaline phosphatase substrate 5-bromo-4-chloro-3-indolyl phosphate.

Enzyme Assays and Kinetics-- NMT activity was assayed as described (34, 46). For the standard enzyme assays, the reaction mixture contained 0.4 µM [³H]myristoyl-CoA, 50 mM Tris-HCl, pH 7.8, 0.5 mM EGTA, 0.1% Triton X-100, synthetic peptide and AtNMT1 enzyme in a total volume of 25 µl. The reaction was initiated by the addition of radiolabeled [³H]myristoyl-CoA and incubated at 30 °C for 10-30 min. The reaction was terminated by spotting aliquots of incubation mixture onto P81 phosphocellulose paper discs and drying them under a stream of warm air. The P81 phosphocellulose paper discs were washed in three changes of 40 mM Tris-HCl, pH 7.3, for 90 min. The radioactivity was quantified in 7.5 ml of Beckman Ready Safe Liquid Scintillation mixture using a Beckman Liquid Scintillation Counter. Kinetic studies were carried out by varying the reaction time, or enzyme and peptide substrate concentrations, or the pH. The utilization of other labeled fatty acyl-CoAs by recombinant AtNMT1 was also examined by replacing [³H]myristoyl-CoA in the standard assay employing synthetic peptides derived from plant and mammalian proteins. One unit of AtNMT1 activity was expressed as 1 nmol of myristoyl-peptide formed per min per mg protein. The AtNMT1 inhibitory assay was carried out using NIP₇₁ according to the method described earlier (36). Protein concentrations were determined by the method of Bradford (47) using Bio-Rad kit and BSA as standard.

Activities of the subcellular marker enzymes were determined according to published methods as follows: succinate dehydrogenase activity for mitochondria and NADPH cytochrome c reductase activity for the microsomes (48); alcohol dehydrogenase activity for the cytosol (49); ADP-glucose pyrophosphorylase activity for the chloroplasts (50). The enrichment of ribonucleoprotein in the ribosomal fraction was assessed by A₂₆₀/A₂₈₀ and A₂₃₅/A₂₈₀ absorption ratios (51).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cloning of an Arabidopsis cDNA Encoding Myristoyl-CoA:Protein N-Myristoyltransferase (AtNMT1) and Its Predicted Structural Features-- To identify NMT homologues in Arabidopsis, we first demonstrated the presence of NMT in Arabidopsis by a series of biochemical experiments. The leaf extracts showed an enzyme activity of 165 ± 6 pmol·min¹·mg¹ protein with pp60^SRC peptide. Additionally, the Western blot analyses indicated that hNMT antibody recognized a prominent immunoreactive band of approximately 50-kDa in Arabidopsis (data not shown). These experiments suggested that Arabidopsis plants show NMT activity and also that the plant NMT had structural similarity to hNMT. Accordingly, a full-length cDNA encoding hNMT1 was used as a probe to screen a cDNA library of Arabidopsis, and three positive clones were obtained. Subsequently, database searches using the conserved domains of yeast and hNMT identified an Arabidopsis partial EST clone (GenBank^TM accession number T21207). This cDNA fragment was used for further screening, and five clones were identified. The eight clones from these screens were sequenced, analyzed, and shown to represent a single gene encoding an NMT homologue (AtNMT1). The sequencing of the largest cDNA revealed a putative 1302-bp ORF flanked by 269 and 139 bp of 5'- and 3'-untranslated regions (UTRs), respectively (GenBank^TM accession number AF193616). This ORF encodes a predicted polypeptide of 434 aa with a calculated molecular mass of 48,706 Da (Fig. 1). A total of four putative translational start sites (ATGs) was found in the long 5'-UTR. The first and second specify very short peptides of 4 and 3 aa, which are unlikely to represent translational products, whereas the third is predicted to encode a 27-aa peptide and has a less favorable translational context surrounding the ATG (CACATGAC). The fourth ATG specifies the longest ORF and is the most likely site of translational initiation due to both the presence of an in frame stop codon 9 bp upstream and to its more favorable translational context (GAAATGGC) (52).

View larger version (130K): [in this window] [in a new window]   Fig. 1.   Analysis of predicted primary structures of orthologous NMTs. The multiple sequence alignment was generated using the CLUSTAL PC/GENE program. The conservation is represented by the following colors: primary (red, 100%), secondary (green, 90%), and tertiary (yellow, 80%). The primary sequences of C. albicans, S. cerevisiae, C. neoformans, H. capsulatum, human, and C. elegans NMTs were taken from Zhang et al. (74). Bovine NMT was taken from Raju et al. (75). Human NMT 1 and 2 and mouse NMT 1 and 2 were taken from Giang and Cravatt (56). Drosophila NMT was taken from Ntwasa et al. (29). Numbers indicate amino acid positions.

Sequence comparison revealed that AtNMT1 showed a high degree of similarity with NMTs from other species at the primary structure level. The predicted AtNMT1 exhibited a 50% identity with human and bovine NMTs, 44% with Drosophila NMT, and 38% with yeast NMT (Fig. 1). A high degree of sequence conservation was also found throughout these primary sequences, including the residues essential for myristoyl CoA binding, the peptide binding, and the active site of the enzyme (53-55). In addition, two signature sequences of NMT, "KFGXGDG" and "(ED)(IV)NFLXHK," are also conserved in AtNMT1 (Fig. 1). Sequence divergence was also found in the amino-terminal domain of the NMTs, the region that likely functions in protein targeting but not in catalytic activity (33, 56). Furthermore, the Lys-rich segment, KKKKKKQKKKKEK, that is conserved in human and mouse NMTs is absent in Arabidopsis, yeast, Drosophila, and Caenorhabditis elegans NMTs (Fig. 1). These positively charged residues have been identified in other proteins involved in the co-translational processing of proteins, including N-methionylaminopeptidase (57, 58).

Arabidopsis Contains Two NMT-like Genes-- The restriction enzymes EcoRI, HindIII, PstI, and XbaI, which do not have recognition sites within AtNMT1 cDNA, were used to digest genomic DNA. These digests all produced two hybridizing bands, indicating that the Arabidopsis genome likely contains two copies of NMT (Fig. 2). The nucleotide sequences of two Arabidopsis genomic clones, AtNMT1-G1 and AtNMT1-G2 matched perfectly with the sequence of AtNMT1 cDNA. Further sequence comparison between the genomic and cDNA clones revealed the presence of one 315-bp intron in the 5'-UTR located three bases upstream to the longest ORF, whereas the 434-aa NMT protein coding region is not interrupted by any introns. The exon/intron junctions have conserved sequences in close agreement with compiled data available for Arabidopsis intron splice sites at the "A. thaliana" database. Further analysis of the Arabidopsis genomic sequence database using our AtNMT1 cDNA and genomic sequences identified two AtNMT1 homologues on different chromosomes. The AtNMT1 cDNA showed 100% identity with a region of a P1 genomic clone (MHM17, GenBank^TM accession number AB024035) (which shows 100% match with AtNMT1-G1 and -G2) on chromosome 5 and 77% identity with BAC F6E13 genomic clone (GenBank^TM accession number ATAC004005) carrying a fragment of chromosome 2. Analysis of BAC F6E13 genomic sequence by The Institute for Genomic Research was predicted to encode a 389-aa NMT-like polypeptide, and we designated this as AtNMT2. We carried out further comparative analysis with all the known NMTs including AtNMT1, and the multiple alignment results show that the predicted AtNMT2 has 31, 27, and 25% identity with human and bovine, Drosophila, and yeast NMTs, respectively (Fig. 1). A single potential intron of 185 bp is located within the protein coding region of putative AtNMT2 (between aa 100 and 101). The intron locations and sizes are significantly different between the two Arabidopsis NMTs. To address the identity of the two hybridizing bands on Southern blot (Fig. 2) in relation to the predicted Arabidopsis genomic NMT sequences, we isolated the restriction fragments corresponding to the size range of two hybridizing bands separately and performed PCR employing primers designed for the conserved protein coding domains of AtNMT1. Sequence analyses of these two independent PCR products revealed that one matches perfectly with the predicted AtNMT1 on chromosome 5, and the other shows 100% match with the putative AtNMT2 on chromosome 2 (data not shown). These results collectively demonstrate that the Arabidopsis contains two organizationally distinct NMT genes that are located on different chromosomes.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Molecular phylogenic tree of the amino acid sequences of NMTs from various species. The tree was constructed by the neighbor joining method, based on sequence information as described in Fig. 1.

Molecular Phylogenic Relationship of AtNMT Proteins with Other NMTs-- Phylogenic analysis of the NMT family reveals that the members can be grouped into two major families (Fig. 3). One family is composed of proteins from S. cerevisiae, Cryptococcus neoformans, C. albicans, and Histoplasma capsulatum, and the other family contains proteins from Arabidopsis, C. elegans, Drosophila, human, mouse, and bovine. These can be further sub-grouped into seven distinct subfamilies based on the conservation of aa between various NMT proteins of divergent origin. These groups include proteins from C. neoformans as subfamily 1; H. capsulatum, S. cerevisiae, and C. albicans as subfamily 2; Arabidopsis as subfamily 3; C. elegans as subfamily 4; Drosophila as subfamily 5; human NMT2 and mouse NMT2 as subfamily 6; and bovine NMT, human NMT1, and mouse NMT1 as subfamily 7. The phylogenic tree suggests that AtNMT1 and AtNMT2 belong to a new subfamily. Despite the divergence of these species over several hundred million years, there is a high degree of sequence conservation of NMT, especially at the aa levels. This high level of conservation suggests some functional implications concerning this gene family.

View larger version (65K): [in this window] [in a new window]   Fig. 3.   Southern blot analysis of Arabidopsis genomic DNA. Genomic DNA (5 µg) was digested with EcoRI, HindIII, PstI, and XbaI, respectively, separated, transferred, and hybridized as described under "Experimental Procedures." The sizes of molecular weight markers in kilobase pairs are marked on the left.

Biochemical Characterization of Recombinant AtNMT1-- To investigate the functionality of the cloned AtNMT1, the gene was expressed in E. coli as a fusion protein with a amino-terminal polyhistidine tag (pTrcHisC-AtNMT1). The specific activity of the purified recombinant AtNMT1 was 1.15 nmol·min¹·mg¹·protein in the presence of pp60^SRC peptide. The protein exhibited an apparent molecular mass of 50 kDa under denaturing conditions (data not shown). The non-transformed or empty vector controls did not have NMT activity. The enzyme was active at a pH range of 5.5-8.5, with a maximum activity at pH 7.8. The metal ions Mg²⁺, Ca²⁺, and Mn²⁺ had negligible effects on AtNMT1 activity (data not shown), indicating that the enzyme does not require these divalent cations for its activity. The addition of 1 mM EDTA and 0.1% Triton X-100 had modest stimulatory effects (18 and 25% increase, respectively) on the activity, compared with the untreated enzyme. However, the AtNMT1 was inhibited by bovine brain NIP₇₁ (36) in a dose-dependent manner (IC₅₀ = 250 ng, Fig. 4). To examine the fatty acyl-CoA chain length donor specificity of AtNMT1, the labeled decanoyl- (C10:0), lauroyl- (12:0), myristoyl- (C14:0), palmitoyl- (C16:0), and stearoyl (C18:0)-CoAs were tested using Arabidopsis AtCDPK6-derived peptide as a substrate. As shown in Fig. 5, the recombinant AtNMT1 exhibited a relative specificity for myristoyl-CoA as the acyl donor. The C10:0 and C12:0 fatty acyl-CoAs were utilized as substrates at 40 and 25% of myristoyl-CoA. However, palmitoyl- and stearoyl-CoA were only negligibly active as acyl donors. Furthermore, the AtNMT1 has a high affinity for myristoyl-CoA (K_m = 3.7 µM) in the presence of AtCDPK6 derived-peptide substrate. These results collectively indicate that the biochemical properties of AtNMT1 are comparable to human and yeast NMTs and, notably, that the fatty acid substrate specificity of this enzyme is highly conserved among these eukaryotic species.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Inhibitory effect of the bovine NMT inhibitor on AtNMT1. The recombinant AtNMT1 (2.5 µg/assay) was incubated with the NIP71 protein (0-600 ng/assay) and measured using pp60SRC-derived peptide as described under "Experimental Procedures."

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Acyl chain specificity of AtNMT1. The specificity of NMT for acyl group was examined using 5 µM acyl-CoA and 50 µM AtCDPK6-derived peptide in the standard assay.

Comparative studies of peptide substrate specificities between AtNMT1 and hNMT1 recombinant enzymes revealed significant differences in their respective catalytic efficiencies (the details are summarized in Table I). The results suggested that generally AtNMT1 had lower K_m values and higher V_max/K_m values toward plant peptides derived from the amino-terminal sequence of Fen kinase and AtCDPK6, whereas these peptides produced higher K_m values and lower V_max/K_m values with hNMT1. The peptide derived from tobacco NtCDPK1 exhibited a 4-fold higher K_m value and a significantly lower V_max/K_m value with AtNMT1. Interestingly, AtNMT1 showed a 12-fold higher K_m and 200-fold lower V_max/K_m values toward peptide derived from tomato DEM as compared with the peptides from Fen kinase and AtCDPK6. However, no detectable activity was observed for this peptide with hNMT1. On the other hand, the results from mammalian peptides suggest that AtNMT1 had 2-fold lower K_m value for pp60^SRC than for the cAMP-dependent protein kinase-derived peptide substrate but with similar V_max. All the four mammalian peptides exhibited significantly lower catalytic efficiency for AtNMT1 compared with hNMT1. Taken together, the results clearly suggest that AtNMT1 and hNMT1 show significant differences in their peptide substrate specificities.

It was also observed that addition of an amino-terminal methionine residue to AtCDPK6-derived peptide (yielding MGHRHSKSKK) or substituting Ala for the Gly^2nd (AHRHSKSKK) resulted in no detectable activity for both AtNMT1 and hNMT1 (Table I), indicating that the amino-terminal glycine residue is essential for recognition by AtNMT1 and hNMT1. This is consistent with results from other species (3).

Expression of NMT in Arabidopsis-- Northern blot analysis with an AtNMT1 cDNA probe identified a single hybridizing band of 1.7 kb in roots, leaves, stems, flowers, and young siliques of Arabidopsis (Fig. 6A). The levels were higher in young roots, leaves, and siliques. To evaluate whether the transcript profile corresponds to protein level, anti-AtNMT1 peptide antibody was used to detect endogenous NMT protein. A single immunoreactive band of 50 kDa was detected in the extracts of leaves, roots, stems, flowers, and siliques (Fig. 6B), and there was a higher level of this protein in young leaves, roots, and siliques. These results were consistent with the Northern blot analyses. Additionally, enzyme assays confirmed the existence of NMT activity in all tissues analyzed and the highest activity in the younger tissues (Fig. 7).

View larger version (42K): [in this window] [in a new window]   Fig. 6.   AtNMT gene expression patterns in various tissues of Arabidopsis plants. A, top panel, RNA blot analysis of steady state AtNMT transcript levels. Bottom panel, ethidium bromide-stained intensity of the ribosomal bands for each RNA preparation. B, relative level of transcripts. The intensity of hybridization band of stem 35 days after planting was set at 1. C, Western blot analysis of AtNMT polypeptide levels. Protein samples (10 µg) from 20,000 × g supernatant of different tissues were run on SDS-polyacrylamide gel for the Western analysis as described under "Experimental Procedures." dap, days after planting. daf, days after flowering.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   AtNMT activity in different tissues of Arabidopsis. The supernatant of extracts from various tissues centrifuged at 20,000 × g was used as enzyme source for AtNMT assay as described under "Experimental Procedures." L, leaf; R, root; St, stem; F, flower; Si, silique. Numbers denote days after planting or flowering.

Subcellular Distribution of AtNMT-- The mitochondrial, chloroplastic, microsomal, ribosomal, and cytosolic supernatant fractions of the Arabidopsis leaf extract from 20-day-old seedlings were examined for NMT activity along with the activities of subcellular marker enzymes as the controls. As shown in Table II, the NMT activity was barely detectable in mitochondrial, chloroplastic, or microsomal fractions. Instead, 58% of the total AtNMT activity was found in the enriched ribonucleoprotein fraction that exhibited the highest A₂₆₀/A₂₈₀ and A₂₃₅/A₂₈₀ absorbance ratios, whereas 33% of the total AtNMT activity was observed in cytosolic fraction. The relatively low level of chloroplastic (1.6%), mitochondrial (0.7%), microsomal (6.4%), and cytosolic (2.8%) marker enzyme activities in the ribosomal fraction rules out the possibility of gross contamination by other major subcellular fractions. In addition, relatively low levels of the other organellar marker enzyme activities were found in the cytosolic fraction, indicating the absence of major contribution by other subcellular organelles to the observed NMT activity in this fraction. Furthermore, a predominant association of NMT protein with the ribosomal fraction was also seen when the subcellular fractions were immunoblotted with anti-AtNMT1 peptide antibody (Fig. 8). The intensity of the immunoreactive band in the ribosomal fraction was over 2 times stronger than that in the cytosolic fraction. A relatively weak band was also observed in the mitochondrial fraction, consistent with a low level (6.6%) of NMT activity in this fraction (Table II). No immunoreactive band was observed in the other subcellular fractions (Fig. 8).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Immunoblotting of AtNMT in subcellular fractions from Arabidopsis leaves. Aliquots (10 µg) of subcellular fractions from Arabidopsis prepared as described under "Experimental Procedure" were analyzed by SDS-polyacrylamide (10%) gel and immunoblotted with an anti-AtNMT peptide antibody. Cyto. + Ribo. indicates 20,000 × g supernatant including cytosolic and ribosomal fractions.

Antisense Suppression of AtNMT1 in Transgenic Arabidopsis Plants Affects Normal Growth-- To determine if the Arabidopsis NMT has any biological significance, we first attempted, without success, to isolate loss of function mutations (null) in AtNMT1 by screening approximately 10,000 T-DNA insertion lines obtained from the Arabidopsis Biological Resource Center. We then generated 16 transgenic lines of Arabidopsis with an antisense AtNMT1 under the control of an enhanced CaMV 35S promoter. In four independent experiments, the AtNMT1 antisense construct yielded significantly lower numbers of transgenic plants. Notably, four of the transgenic plants recovered from kanamycin selection did not survive long enough to produce seeds. These results suggest detrimental effects of the suppression of NMT activity in some lines. We need additional evidence for this. The 12 kanamycin-resistant lines that survived were confirmed to be transgenic by PCR and Southern blot analyses (data not shown). We analyzed the T2 generation of the transgenics for growth and morphology. Out of these 12, 2 lines (Anti-AtNMT1-1 and Anti-AtNMT1-2) exhibited a significant reduction in the growth and displayed fewer rosette leaves (Fig. 9A), whereas the phenotype in other lines were close to wild type. More detailed studies were carried out with anti-AtNMT1-1 and anti-AtNMT1-2 lines. Overall these plants were smaller (dwarf) compared with wild type. The two lines that showed severe growth phenotype also displayed a 2-3-fold lower biomass in comparison with wild type (Fig. 9B). The AtNMT1 transcript and protein levels were significantly reduced (70-90%) in both these lines (Fig. 9, C and D), showing strong correlation between the NMT suppression and growth defects.

View larger version (30K): [in this window] [in a new window]   Fig. 9.   Effects of antisense suppression on plant growth. A, the antisense AtNMT1 transgenic plants exhibited abnormal morphology. WT, wild type. Anti-NMT1-1 and -2, antisense AtNMT1 transgenic line 1 and 2, respectively. B, vegetative growth characteristics in control and antisense transgenic T2 plants. Data presented are the means ± S.E. for 5 plants from each line. WT, wild type. C, top panel, RNA blot analysis of AtNMT transcript levels. Samples for RNA extraction were harvested from control and antisense transgenic plants (20 days). Bottom panel, ethidium bromide staining intensity of the ribosomal bands for each RNA preparation. D, Western blot analysis of AtNMT polypeptide levels in control and the transgenic plants. The method was as described in Fig. 6.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

N-Myristoylation of proteins is a co-translational event (3, 4, 59), and it appears to play an important role in mediating specific protein-protein or protein-lipid interactions, ligand-induced protein conformational changes, and subcellular targeting (1, 3, 10, 60). The enzyme responsible for this modification, NMT, has been extensively studied in yeast and mammalian cells in the past decade, partly because NMT is a potential target for antiviral, antifungal, and antineoplastic therapy (1, 2). However, very little information is available about myristoylation in plants. We have now demonstrated NMT activity in Arabidopsis and characterized it further. Given the wealth of information accruing in Arabidopsis, this work provides an excellent means to study the functions of NMT in plants.

A full-length cDNA isolated from Arabidopsis was confirmed to encode an NMT by its enzymatic activity in a heterologous expression system (Table I). The primary structure of AtNMT1 generally resembles that of mammalian NMTs in that it shows a high similarity with these NMTs and possesses almost all amino acid residues essential for catalytic activity. The hNMT antibody could recognize Arabidopsis NMT, further confirming its overall structural similarity. However, a sequence divergence has been observed in the amino-terminal non-catalytic region of these NMTs. This is consistent with the suggestion that the amino-terminal domains of NMTs have regulatory functions, thus necessitating species-specific requirements (33, 61, 62). Although the precise role remains to be defined, the amino terminus of NMT has been implicated in mediating protein-protein or protein-lipid interactions, which determine targeting to specific intracellular compartments and regulate co-substrate availability through associations with intracellular proteins (29, 30, 33). However, the role of the amino terminus of AtNMT1 remains to be defined.

The predicted molecular mass of AtNMT1, 48,706 Da, is in good agreement with the E. coli-expressed recombinant protein (data not shown) and the leaf AtNMT1 (Fig. 6C), which was estimated to be 50,000 Da. This is in general agreement with previous studies of other NMTs that gave molecular mass of 50-60 kDa for monomeric human (26, 33), 50 kDa for bovine spleen (28) and cardiac muscle (63), 55 kDa for yeast (64), 53 kDa for C. albicans (25), and 46 kDa for Drosophila (29). However, NMTs from murine leukemia cell line L1210 (65) and bovine brain (27) have been demonstrated to exist in multiple isoforms.

The existence of two AtNMT genes in Arabidopsis adds to the complexity of AtNMT function in plants. Sequence analysis of two genomic clones (AtNMT1-G1 and -G2) revealed a perfect match with AtNMT1 cDNA, and database searches identified the second putative AtNMT2 genomic region from chromosome 2. Furthermore, it is interesting to note that the only intron in AtNMT1-G1and G2 on chromosome 5 is located 3 bp upstream to the longest ORF of AtNMT1, whereas the single intron predicted in AtNMT2 gene on chromosome 2 is located in the protein coding region after the 100th amino acid. In addition, an alignment of these two sequences showed that the two proteins share 71% similarity with the most divergence in amino-terminal domains and displayed no similarity at all between positions 73 and 145 of AtNMT1 (Fig. 1). It should also be noted that the NMT signature sequences (KFGXGDG and (ED)(IV)NFLXHK) were not conserved in the putative AtNMT2, although it showed conservation in residues, such as Cys-169, His-293, and Gly-412 from human and bovine NMTs, which have been implicated in the catalytic activity (Fig. 1). These observations further indicated that Arabidopsis contains two potential AtNMT genes with distinct genomic organization. These data suggest that AtNMT1 and AtNMt2 must have diverged early or the conservation shows a convergent evolution of related functions. In a separate study, we found two copies of NMT-like genes in diploid and four copies in tetraploid Brassica species,² consistent with the Arabidopsis results. The function of the AtNMT2 gene is unknown. A more precise investigation of the functional differences, if any, between the two AtNMT genes in Arabidopsis can be pursued when the full-length AtNMT2 cDNA is isolated. Interestingly, mammals also contain two NMT genes, human-NMT1 and NMT2 and mouse-NMT1 and NMT2 (56). Their genomic organization is complex. For example, hNMT1 is encoded by 11 exons of various lengths (66). Furthermore, as with AtNMTs, aa sequence divergence was most prevalent in the amino-terminal domains (67).

While the transcript, polypeptide, and enzyme activity of AtNMT were detected in all Arabidopsis tissues analyzed, the greater level in young roots, leaves, and siliques suggest a developmental regulation. However, a more precise investigation of spatial distribution of AtNMT expression is required for this conclusion. It should be noted that the transcripts and enzyme activity measured in Arabidopsis tissues may come from both AtNMT1 and AtNMT2, because they are highly homologous and could utilize similar peptide substrates. However, the observed variation in polypeptide levels (Fig. 6C) only reflects the difference in level of AtNMT1 expression among these different stage tissues, since the anti-AtNMT1 peptide antibody used was produced based on the peptide from aa 114 to 129 of AtNMT1, which are completely absent in putative AtNMT2 aa sequence (Fig. 1). The human and bovine NMTs are also expressed in all the tissues tested, with variation in the levels of transcripts and activities (34, 68). In Drosophila, a higher level of NMT transcript occurs at early stages of embryogenesis (29). Thus there appears to be a developmental regulation of NMTs.

AtNMT1 has similar biochemical properties with other eukaryotic NMTs in terms of optimum pH and fatty acid chain-length specificities. As shown in yeast (69) and human (70) NMTs, the recombinant AtNMT1 had a high degree of specificity for myristoyl-CoA as acyl-CoA donor. The palmitoyl- and stearoyl-CoAs did not function as a substrate, whereas shorter chain (C10:0 and C12:0) acyl-CoAs donated their acyl chains at significantly reduced rate (25-40% of myristoyl-CoA). The AtNMT1 specificity for myristoyl-CoA (a less hydrophobic acyl chain in comparison with long chain fatty acids) is significant in light of the higher availability of other fatty acids in plant tissues, for example, the pool size of palmitic acid (C16:0) is 20 times that of myristic acid and more than 100 times that of decanoic and dodecanoic acids in Arabidopsis Columbia leaves,³ and yet these do not function as the acyl donors. A similar inference has been made in other eukaryotes (3). In vitro kinetic studies also revealed that the enzyme exhibited a great deal of specificity for its peptide substrate and recognized only those with the amino-terminal glycine residues (Table I). However, AtNMT1 has a large variation in its catalytic efficiency (V_max/K_m) for various peptide substrates (Table I), indicating preference among these peptides and suggesting that proper context in the polypeptide substrates (besides conserved glycine and serine residues at 2 and 6, respectively) may play an important role. Furthermore, comparison of the peptide substrate specificities of AtNMT1 and hNMT1 using synthetic peptides derived from the amino-terminal sequences of eight different N-myristoyl proteins indicated that these two NMTs displayed significant differences in their catalytic efficiencies (V_max/K_m). For instance, the two peptides derived from M2 and myristoylated alanine-rich protein kinase C substrate were utilized readily by hNMT1 but not for AtNMT1. Conversely, the two peptides derived from Arabidopsis AtCDPK6 and tomato Fen kinase were more efficiently myristoylated by AtNMT1 in comparison to hNMT1 (7-15-fold greater catalytic efficiency, see Table I). In general, the peptides derived from the putative plant N-myristoylated proteins gave a higher catalytic efficiency than those from the mammalian proteins. Similar comparative studies using NMTs from the yeast, rat liver, and wheat germ showed significant differences in terms of their respective peptide substrate specificity (10, 67, 71). The DEM gene product of tomato plays an important role in meristem function, and it contains potential N-myristoylation signal (24). However a peptide designed for it was not efficiently myristoylated by AtNMT1. This peptide contained a negatively charged (glutamic acid) residue at position 8, and this might have interfered with the modification as reported in other studies (61, 69, 70). Compared with the DEM, the peptide derived from tomato Fen kinase exhibited high catalytic efficiency with AtNMT1. These results suggest that Fen kinase is myristoylated in vivo and further support the proposed functional role for the amino-terminal Gly-2 residue of this enzyme (72). Analysis of DNA and protein databases for potential N-myristoylation sites in plant gene products revealed CDPK members as the largest group that contains this modification signal. In the present study, the peptides derived from Arabidopsis and tobacco CDPKs served as substrates for AtNMT1, and these results further suggest that this modification likely play an important in vivo biological role for these proteins.

Our observations indicate an association of the majority of AtNMT with the subcellular fraction enriched in ribosomes (Table II and Fig. 8). The mechanism by which AtNMT is associated with the ribosomal fraction is unknown at present. A careful examination of hNMT subcellular localization established an apparent specific association of hNMT activity and polypeptide with the subcellular ribosomal fractions from human lymphoblastic leukemia and cervical carcinoma cells (33). It was further proposed that the amino-terminal domain of hNMT could provide a ribosomal targeting signal for N-myristoylation and be involved in an interaction of hNMT with other factors on ribosomes (33). Drosophila NMT has been shown to be associated with the membrane by interaction with p34 ribosome-binding protein (29). Analysis of AtNMT1 sequence indicated the presence of potential transmembrane regions (data not shown). These residues may provide a targeting signal to the site of ribosomes. In addition, the recovery of AtNMT activity in the ribosomal fraction may be underestimated because of membrane-associated NMT inhibitor(s). NMT inhibitor activities have been observed in yeast (70), rat (67), and bovine brain (36) tissues. A membrane associated 71-kDa NMT inhibitor protein (NIP₇₁) has been purified and characterized from bovine brain (36). This protein was found to inhibit the recombinant AtNMT1 activity in a dose-dependent manner with a half-maximal inhibition of 250 ng. Furthermore, a strong immunoreactive band was also detected in the ribosomal fraction of Arabidopsis leaf extracts probed with an antibody directed against NIP₇₁ (data not shown). It may be possible to speculate that this inhibitor protein could serve as a regulator of several eukaryotic NMT activities. Further detailed studies, however, are required to understand the mechanism of regulation of NMT activity in the ribosomal subcellular fractions.

To address the biological function for AtNMT1 in Arabidopsis, we took an antisense transgenic approach. The present study demonstrated that the suppression of AtNMT1 expression had a significant effect on the plant growth. Obviously, if NMT is essential to the existence of the plant as it is for yeast (8), then only those escaping a severe suppression will be selected in the course of genetic transformation. Thus, among the 12 plants, two showed a severe reduction in the growth. Since the empty vector control did not exert a negative effect, the stunted growth of the transgenics is likely due to an impairment of myristoylation. These results also suggest that the low NMT expression levels likely contributed as a limiting factor for loss of myristoylation-dependent biological functions in some potential protein targets. These potential targets may include serine/threonine kinases, tyrosine kinases,  subunit of heterotrimeric G-proteins, or other yet to be identified proteins that contain putative N-myristoylation consensus motif. These proteins participate in signal transduction pathways and developmental processes (14-16, 18, 19, 21), and thus the AtNMT1 antisense phenotype is conceivably a global effect of mis-regulating these proteins. Interestingly, the AtNMT1 antisense phenotype in some respects is reminiscent of recently described dwarf plant phenotype in transgenic rice plants produced by antisense suppression of the heterotrimeric G-protein  subunit (73). The Arabidopsis  subunit gene product also contains a potential N-myristoylation site (21). In a separate in vivo E. coli co-expression study, we found that rice  subunit served as an efficient substrate for N-myristoylation by AtNMT1.³ It is possible that the observed growth effects in transgenic Arabidopsis could be partly mediated by reduced  subunit functional activity. However, it is clearly evident from the antisense transgenic results that AtNMT1 has a functional role in normal growth and development.

In conclusion, we have shown that a functional NMT is present in all organs of Arabidopsis and that AtNMT1 differs from other NMTs with respect to its peptide substrate preference. The present study also showed that AtNMT has an important biological function in plants. The availability of AtNMT gene, recombinant NMT protein, antibody, and information regarding AtNMT protein properties will facilitate further investigations to determine the role(s) of protein N-myristoylation as well as to define NMT protein targets in plants.

	ACKNOWLEDGEMENTS

We thank the Alberta Peptide Institute, Canada, for synthesis of peptides, and Don Schwab, Barry Panchuk, and Dr. Larry Pelcher of the DNA Technology Group of PBI for primer synthesis and DNA sequencing. We thank Arabidopsis Biological Resource Center (Ohio) for providing the EST clone T21207. We also thank Drs. Adrian Cutler, Pierre Fobert, Jitao Zou, ASN Reddy, Tim Dumonceaux, and Prakash Venglat for their helpful comments and suggestions in the preparation of this manuscript.

	FOOTNOTES

^* This is publication number 42625 from the National Research Council of Canada.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF193616.

^§ Both authors contributed equally to this work.

Current address: Agriculture & Agri-Food Canada, Saskatoon Research Center, Saskatoon S7N 0X2, Saskatchewan, Canada.

^** To whom correspondence should be addressed: National Research Council of Canada, Plant Biotechnology Institute, 110 Gymnasium Place, Saskatoon S7N 0W9, Saskatchewan, Canada. Tel.: 306-975-5267; Fax: 306-975-4839; E-mail: raju.datla@pbi.nrc.ca.

² C. Jiang, Q. Qi, G. Selvaraj, and R. Datla, unpublished results.

³ Q. Qi and R. Datla, unpublished results.

	ABBREVIATIONS

The abbreviations used are: NMT, myristoyl-CoA:protein N-myristoyltransferase; kb, kilobase pair; bp, base pair; CDPKs, calcium-dependent protein kinases; PMSF, phenylmethylsulfonyl fluoride; h, human; At, A. thaliana; KLH, keyhole limpet hemocyanin; BSA, bovine serum albumin; ORF, open reading frame; UTRs, untranslated regions; aa, amino acids; PCR, polymerase chain reaction.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES