Index INTRODUCTION MATERIALS AND METHODS Plant material. Reverse transcription-polymerase chain reaction (RT-PCR). DNA sequencing. Measurement of RNA accumulation by Northern blot analysis and RT-PCR. Expression of RsNdr1 and RsNdr2 in E. coli. Protein kinase assay. RESULTS Primary structures of RsNdr1 and RsNdr2. Accumulation of RsNdrs mRNA in a plant. Detection of the protein kinase activity of RsNdrs expressed in E. coli. DISCUSSION RsNdrs are plant orthologues of the fungal and animal protein kinases which control cell morphology and cell polarity. Expression of RsNdrs. Protein kinase activity of RsNdr proteins expressed in E. coli. References

INTRODUCTION

Reversible protein phosphorylation, which is catalyzed by protein kinases and phosphatases, plays a crucial role in the regulation of cell metabolism, growth and development. In higher plants, protein kinases function in signal transduction cascades in response to a variety of signals such as light, water, temperature and hormones. Since the pioneering work of Lawton et al. (1989), in which the first plant protein kinase, PVPK-1, was documented, several protein kinases have been isolated, and in some cases, their crucial roles in signal transduction cascades have been clarified (e.g. Meskiene and Hirt 2000, Innes 2001, Asai et al. 2002).

Eukaryotic protein kinases contain 12 highly conserved regions which are referred to as subdomains (Hanks et al. 1988, Hanks and Quinn 1991). Invariant or nearly invariant residues in the subdomains are important for maintaining the three-dimensional structure of the proteins and their catalytic activity. Based on a phylogenetic analysis of the amino acid sequences of the subdomains, Hanks and Hunter (1995) first classified the eukaryotic protein kinases into the five main groups: the ‘AGC’, ‘CaMK’, ‘CMGC’, ‘PTK’ and ‘other’ groups. After that, Stone and Walker (1995) applied this classification with some modifications for plant protein kinases, and pointed out the similarities and the differences between plant protein kinases and other eukaryotic protein kinases. Among plant protein kinases, those belonging to the AGC group are poorly characterized, except for a few cases (Zhang et al. 1994a, b). The AGC group is represented by the cyclic nucleotide-dependent kinases (PKA and PKG), the calcium-phospholipid-dependent kinases (PKC) and the ribosomal S6 protein kinases. These kinases share a common feature: they are regulated by second messengers (e.g., cAMP, cGMP, diacylglycerol or Ca²⁺), with the exception of the ribosomal S6 protein kinases. No authentic PKA has ever been cloned from any plant species, and plants use Ca²⁺ signaling in a different way from mammals. The major calcium-dependent kinases cloned thus far are CDPKs, which are structurally different from PKC and unique to plants. The CDPKs belong to the CaMK group. Considering the importance of cAMP and Ca²⁺ as ubiquitous messengers, the absence of PKA and PKC, and the absence of signal transduction cascades in which both protein kinases are involved, are somewhat puzzling, and we therefore conducted a search for protein kinases in the AGC group from higher plants.

Radish is an annual vegetable in the genus Brassicaceae and is widely cultivated in Japan. Besides its long history of use in the field of plant physiology, radish is a useful source for isolating novel genes due to the tremendous amount of information available on the model plant Arabidopsis thaliana. In addition, because the plant size of radish is larger than that of Arabidopsis, studying radishes rather than Arabidopsis is advantageous for a variety of biochemical methods. In this study, in order to gain a better understanding of plant protein kinases in the ‘AGC’ group, we cloned and characterized three cDNAs encoding novel protein kinases, RsNdrs, and expressed them in E. coli to examine their enzymatic activities. The results show that at least one of the RsNdrs is functional and that RsNdrs are a new family of protein kinases in radish.

MATERIALS AND METHODS

Plant material.

A European cultivar of radish (Raphanus sativus L. cv. ‘Comet’) was used as the material plant. The plants were grown to maturity in a greenhouse under normal day-light/temperature conditions.

Reverse transcription-polymerase chain reaction (RT-PCR).

Total RNA was isolated from 100 mg of young leaves using an RNeasy plant mini kit (Qiagen). cDNA was reverse-transcribed from 5 μg of total RNA using the Superscript preamplification system for first-strand cDNA synthesis (GIBCO BRL). PCR amplification was carried out on a tenfold dilution of the RNaseH-treated cDNA mixture with the following four oligonucleotide primers (forward: #1; 5’-AGAGGACAGGTTGAGCATGT-3’, #2; 5’-ATATGATGACCTTGCTCATGAG-3’, reverse: #3; 5’-GAGGATACCCAACGAGCATTTC-3’, #4; 5’-CATGCCATATCCTTTCTTCAGC-3’). The primers were designed based on the nucleotide sequences of Arabidopsis ESTs (N64988, nucleotide positions 134~153 and 270~291, and N65509, nucleotide positions 177~198 and 119~140), referring to the sequences for the catalytic domains of PKAs from budding yeast (GenBank accessions M17072-04, M17224 and Y00694) and fission yeast (U08622), and a cDNA for a putative protein kinase from tobacco (=NTPK7, X71057). PCR was performed in a System 2400 programmable incubator (PE biosystems) using ExTaq DNA polymerase (Takara) with the following settings: initial denaturation (94°C, 3 min), 30 cycles of denaturation (94°C, 30 sec), annealing (55°C, 30 sec) and extension (72°C, 1 min). The PCR products were cloned into a plasmid vector, pGEM-T Easy (Promega), and their nucleotide sequences were determined as described below. Rapid amplification of cDNA ends (RACE) was performed with various radish-specific primers designed based on the DNA sequences of the initial PCR products. Primers specific for the 5’ and 3’ ends of the radish cDNAs were designed based on the sequences of the RACE products, and then putative full-length cDNAs were amplified by RT-PCR. Complete nucleotide sequences of the putative full-length cDNAs were determined after the cDNAs were cloned into a plasmid vector as described below. The sequences were deposited in the DDBJ database under accession numbers AB105045-AB105047

DNA sequencing.

The nucleotide sequences of the selected plasmid DNAs were determined by the cycle sequencing procedure using an ALFexpress DNA sequ-encer (AP Biotech) following the instructions of the manufacturer. The nucleotide and amino acid sequences were analyzed using the software GeneWorks (Oxford Molecular Group), clustal W at Genomenet (http://clustalw.genome.ad.jp/), PAUP* version 4.0 (Swofford, 2002), and the BLAST program at NCBI (http://www.ncbi.nlm. nih.gov/BLAST/).

Measurement of RNA accumulation by Northern blot analysis and RT-PCR.

For Northern blot analysis, total RNA was prepared from the mature and young (6-week old) leaves, stems, roots and flower buds of the plants using Sepasol-RNAI (Nacalai Tesque). About 50 μg of total RNA was electrophoresed on a 1.2% agarose gel containing 5.0% (v/v) formaldehyde and blotted onto a nylon membrane, Biodyne Plus (Pall). The RNA was hybridized with a Dig-labeled RNA probe in Dig Easy Hyb solution (Roche) at 65°C overnight. The Dig-labeled anti-sense RNA probes were prepared by in vitro transcription of the linearized plasmid containing a full-length cDNA for RsNdr1 or RsNdr2. Post-hybridization washes were performed twice at low stringency (2xSSC/0.1%SDS, at room temperature) for 5 min and twice at high stringency (0.1xSSC/0.1%SDS, at 65°C) for 15 min. After the post-hybridization washes, signals were detected by the Dig chemiluminescence detection procedures following the instructions of the manufacturer (Roche). An RNA probe, which was transcribed from a cDNA clone containing radish cytosolic glyceraldehyde-3-phosphate dehydrogenase (GapC-1), was used as a control in the Northern hybridization experiments.

To investigate the accumulation of small quantities of mRNAs, 25 or 30 cycles of RT-PCR were performed with the primer pairs specific for RsNdr1 or RsNdr2. The same RNA as that used in the Northern blot analysis was employed as the template for RT-PCR, except that the RNA was treated with RNase-free DNase to remove residual amounts of contaminaing DNA. GapC-1 was used as a control for RT-PCR experiments.

Expression of RsNdr1 and RsNdr2 in E. coli.

A 1.7-kb SalI cDNA fragment of RsNdr1 and a 1.7-kb BamHI/SalI cDNA fragment of RsNdr2b were cloned into the SalI and BamHI/SalI sites, respectively, of an expression vector pMAL-c2x (NEB). These restriction fragments were obtained by PCR with KOD DNA polymerase (Toyobo) and specific primers in which appropriate restriction sites were incorporated. The resultant fusion constructs were named pMAL-Ndr1 and pMAL-Ndr2, respectively. A fusion construct with a kinase-dead RsNdr1 (K77A) was constructed by PCR-based site-directed mutagenesis of the pMAL-Ndr1 template using Pfu DNA polymerase (Stratagene) and primers #375; 5’-GGCAAACTGAAAAA-AACTG-3’ and #374; 5’-ATGGCATAAACCTCGCCTGT-3’ following the instructions of the enzyme’s supplier (Stratagene).

For expression analysis, an overnight culture of E. coli harboring a fusion construct was transferred to a flask containing 25 ml of SOB, and the E. coli was grown at 37°C with vigorous shaking. When the OD₆₀₀ of the culture reached 0.4~0.6, isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM. One-milliliter aliquots of the culture were taken at 1-hour intervals, and the cells were harvested by centrifugation. The cells were resuspended in 100 μl of 20 mM phosphate buffer (pH 7.0) and disrupted by repeated freeze/thawing. After centrifugation at 14,000 × g for 20 min at 4°C, the supernatant (hereafter termed ‘soluble fraction’) was mixed with an equal volume of SDS sample buffer (0.125 M Tris-HCl pH 6.8, 0.14 M SDS, 20% glycerol, 2% β-mercaptoethanol and bromophenol blue). The remaining pellets (‘insoluble fraction’) were also resuspended in 100 μl of SDS sample buffer. Proteins in the soluble (20 μl) and insoluble fraction (10 μl) were separated by 7.5% SDS-PAGE (Laemmli, 1970), and then stained with CBB. The proteins were transferred onto a nitrocellulose membrane (Hybond ECL; AP Biotech), using a Trans-Blot SD (BioRad). Recombinant fusion proteins were detected using a 1:10,000 dilution of anti-MBP rabbit serum (NEB), HRP-linked anti-rabbit antibodies (Wako), and ECL Western blotting detection reagents (AP Biotech).

Protein kinase assay.

Protein kinase activity was measured using E. coli lysate as an enzymatic source, and histone, myelin basic protein and casein (Sigma) as substrates. The 100-μl reaction contained 1 or 2 μg of proteins from E. coli cell lysate, 20 mM Tris-HCl pH 7.4, 10 mM MgCl₂, 1 mM DTT, 100 mM [γ-³²P]ATP (10 Ci/ml), and 50 μg each of the exogenous substrate proteins. The effects of activating factors on the activity were measured by adding either Ca²⁺ (1.5 mM final concentration) or cAMP (0.01 mM final concentration), or both. When cAMP was used as the sole activating factor, EGTA and EDTA (1 mM each) were included in the reaction mixture. After 2 hours of incubation at 30°C, the reaction was terminated by adding 10 μl of glacial acetic acid, and then 40 μl of the reaction mixture was taken and spotted onto P81 phosphocellulose paper (Whatman). The paper was washed in 75 mM phosphoric acid five times for 5 min each and once in distilled water. The papers were dried and the Cerenkov radiation was measured with a liquid scintillation counter (Packard TRI-CARB 2900TR). As a positive control for the assay, cell lysate prepared from mammalian COS7 cells was used.

RESULTS

Primary structures of RsNdr1 and RsNdr2.

RT-PCR was performed on radish total RNA with the four primers #1–#4, which were designed to amplify a segment between subdomains III and IX in the catalytic domain of PKAs from various organisms. The size of the RT-PCR products which were amplified from the cDNA of leaves was about 450–600 bp with all possible primer combinations, which was slightly larger than the size expected for yeast PKAs (data not shown). DNA sequencing of the cloned PCR products obtained using primer pairs #1/#3 and #2/#4 showed that the former PCR product consisted of two different cDNAs of 583 bp and 601 bp, whereas the latter product was a single cDNA fragment of 486 bp. DNA sequencing of the 3’-RACE product obtained with primer #2 also revealed that this product was a new cDNA fragment different from the aforementioned cDNAs. The deduced amino acid sequences of each of the four cDNA clones contained highly conserved motifs that corresponded to the catalytic domains of eukaryotic Ser/Thr protein kinases. The putative protein kinase revealed by the 3’ RACE-PCR was designated RsNdr1, while the other three protein kinases represented by RT-PCR products of 583 bp, 601 bp and 486 bp were named RsNdr2, RsNdr3 and RsNdr4, respectively. In the rest of this study, RsNdr1 and RsNdr2 were characterized further.

After a series of RACE experiments, primers specific to the 5’- and 3’-ends of RsNdr1 and RsNdr2 cDNAs were synthesized, and a putative full-length cDNA was amplified by RT-PCR. The sizes of the RT-PCR products were about 1.6 kb for RsNdr1 and about 1.8 kb for RsNdr2. DNA sequencing of the 1.6 kb-product showed that the cDNA for RsNdr1 (1643 bp) contained a single open reading frame (ORF) of 1386 bp encoding a protein of 461 amino acids with a calculated molecular mass of 53 kDa (Fig. 1). DNA sequencing of the 1.8-kb product revealed that the product contained two cDNAs of 1730 bp (named RsNdr2a) and 1718 bp (RsNdr2b). RsNdr2a contained a single ORF of 1626 bp encoding a protein of 541 amino acids with a calculated molecular mass of 63 kDa, whereas RsNdr2b contained an ORF of 1557 bp encoding a 518-amino acid protein of 60 kDa (Fig. 2). Although the overall DNA sequences for RsNdr2a and RsNdr2b are quite similar, they differed from each other in the number of GA repeats (10 vs. six in RsNdr2a and RsNdr2b, respectively) at the 5’-end, the 28 synonymous and four nonsynonymous changes in the two ORFs, and the five base changes in the 3’-UTR. There were three insertions/deletions in the 3’-UTR of the two cDNAs (Fig. 2). The deduced amino acid sequences of RsNdr1, RsNdr2a and RsNdr2b contained all 12 of the highly conserved subdomains of the eukaryotic Ser/Thr protein kinases, including the ATP-binding site (in subdomain I-II) and Ser/Thr protein kinase active-site (in subdomain VI) .

View Details

Fig. 1. cDNA (upper) and deduced amino acid (lower, one letter code) sequences of RsNdr1. Conserved catalytic subdomains are shown by bold underlines with Roman numerals. The ATP-binding site and Ser/Thr protein kinase active-site are shown by a shaded and obliquely striped boxes, respectively. The lysine changed to alanine in kinase-dead RsNdr1 (K77A) is boxed. The primers used for amplification of the full-length cDNA are indicated by double underlines.

View Details

Fig. 2. cDNA (upper) and deduced amino acid (lower, one letter code) sequences of RsNdr2a and RsNdr2b. Nucleotides that are identical in RsNdr2b and RsNdr2a are shown by periods (.). Where the two amino acid sequences are identical, only the amino acids for RsNdr2a are shown, and the different amino acids in RsNdr2b are indicated by a one letter code after the slash (/). The putative initiation codon of RsNdr2b is boxed. Gaps are noted with dashes. Conserved catalytic subdomains are shown by bold underlines with Roman numerals. The ATP-binding site and Ser/Thr protein kinase active-site are shown by a shaded and obliquely striped boxes, respectively. The primers used for the amplification of the full-length cDNA are indicated by double underlines, while the anti-sense primer used for RT-PCR is underlined. GA repeats found at the 5’-end of RsNdr2a and RsNdr2b are indicated by a dotted line.

Accumulation of RsNdrs mRNA in a plant.

To study the steady-state level of the RNA accumulation of RsNdrs, Northern blot analysis was conducted with RsNdr1 and RsNdr2b probes using the total RNAs prepared from leaves, stems, roots and flower buds (Fig. 3a). Although a transcript of 1.2 kb was detected clearly with the control probe (GapC), very weak signals of either a 1.8 kb-transcript in stems and flower buds or a 1.7 kb-transcript in roots and young leaves were observed with the RsNdr1 probe (Fig. 3a). No hybridization signal was detected with the RsNdr2b probe under our experimental conditions (Fig. 3a). The accumulation of RsNdrs mRNA was also checked by RT-PCR on the same RNA sample as that used in the Northern blot analysis (Fig. 3b). Using a GapC-specific primer pair, a large amount of RT-PCR product of the expected size was amplified, even after 25 cycles of the reaction. With a primer pair specific for the RsNdr1, a single PCR product of the expected size was amplified from all RNA preparations after 25 cycles of the reaction, while nonspecific minor fragments began to appear after 30 cycles of the reaction in all but root RNA (Fig. 3b). With a primer pair common to RsNdr2a and RsNdr2b, a single PCR product of the expected size was clearly detected after 30 cycles of the reaction (Fig. 3b).

View Details

Fig. 3. Studies of the RNA accumulation of RsNdrs mRNA by Northern blotting (a) and RT-PCR (b). In (a), total RNA prepared from leaves (1), stems (2), roots (3), flower buds (4) and young leaves (5) was hybridized with Dig-labeled GapC (upper), RsNdr1 (middle) and RsNdr2 (lower) RNA probes. The upper part of each figure shows the hybridization signals, while the lower shows the EtBr-stained rRNAs. In (b), RT-PCR products of GapC (upper), RsNdr1 (middle) and RsNdr2 (lower) are shown, together with the number of cycles (25 or 30) employed. Lane 6 shows a negative control of the RT-PCR (without template).

Detection of the protein kinase activity of RsNdrs expressed in E. coli.

To examine whether RsNdr proteins have protein kinase activity in vitro, the putative full-length cDNAs for RsNdr1 and RsNdr2a were each cloned into the expression vector pMAL-c2x and expressed in E. coli as a fusion protein with a tag, maltose binding protein (MBP). A protein of about 100 kDa was overproduced in E. coli harboring pMAL-Ndr1, and was mainly accumulated in the insoluble fraction. Using anti-MBP serum, the 100-kDa protein was immunologically detected in the insoluble fraction together with sev-eral smaller proteins. In the soluble fraction, in contrast, only the 100-kDa protein was detected using the same anti-MBP serum. In the total protein of E. coli harboring pMAL-Ndr2, the overexpressed 110-kDa protein was detected only in the insoluble fraction. Anti-MBP serum detected the 110-kDa protein in both the insoluble and soluble fractions, but the signals were weak, especially in the soluble fraction. In the soluble fraction, several proteins smaller (less than 50kDa) than MBP were detected.

When the cell lysate expressing RsNdr1 fusion protein was incubated with the conventional protein substrates, the highest activity was observed toward myelin basic protein, and weak activity was detected toward the histone. When the activity was measured with the kinase-dead RsNdr1 (K77A), the activity toward both myelin basic protein and histone was lost (Fig. 4). Protein kinase activity was not detected for the RsNdr2 fusion protein using any of three substrates tested (data not shown).

View Details

Fig. 4. Phosphorylation activity of the RsNdr1 fusion protein. A lysate of E. coli cells that overexpressed MBP (pMAL-c2x), RsNdr1 or RsNdr1(K77A) fusion protein (2 μg) was incubated in the kinase buffer containing [γ-32P] ATP, 0.01 mM cAMP, 1.5 mM CaCl2, and each protein substrate. The lysate from mammalian COS7 cells was used as a positive control for the experiment. Bars represent mean +/– standard deviation of duplicate determinations.

DISCUSSION

RsNdrs are plant orthologues of the fungal and animal protein kinases which control cell morphology and cell polarity.

In this study, we cloned and sequenced three putative full-length cDNAs (RsNdr1, RsNdr2a/b) encoding novel protein kinases from radish. The deduced amino acid sequences of all three cDNAs contained all 12 conserved subdomains necessary for the enzymatic activity of the eukaryotic Ser/Thr protein kinase, suggesting that the translation products of these cDNAs are functional. Although the differences between RsNdr2a and RsNdr2b are so small that we can’t assign these two as separate genes (i.e., their relationship might be allelic), cDNA sequencing and Southern blotting analysis (data not shown) showed that at least two different but related protein kinases, RsNdr1 and RsNdr2, are present in radish. BLAST homology search using the deduced amino acid sequences of RsNdrs hit several protein kinases belonging to the AGC group, as expected from our cloning strategy. The AGC group of protein kinases was first defined by Hanks and Hunter (1995), and is represented by PKA, PKG, PKC and ribosomal S6 kinase. Later, the group was divided into nine subgroups based on sequence similarities (Smith et al. 1997). To clarify the relationship of RsNdrs to the protein kinases in the AGC group, molecular phylogenetic analysis was carried out. Fig. 5 shows the phylogenetic tree constructed for 17 protein kinases from various organisms. The tree was reconstructed by the NJ method after aligning the amino acid sequences of the catalytic domains of each protein kinase with the computer program Clustal W. The tree clearly shows that the RsNdrs are closely related to a particular subgroup of the ‘AGC’ protein kinase, consisting of the protein kinases from fungi, i. e., cot-1 (Yarden et al. 1992), TB3 (Buhr et al. 1996), orb6 (Verde et al. 1998), cbk1 (Bidlingmaier et al. 2001) and ukc1 (Durrenberger and Kronstad 1999) of Neurospora crassa, Colletotrichum trifolii, S. pombe, S. cerevisiae and Ustilago maydis, respectively, and those referred to as Ndr (named after nuclear dbf related) identified in humans, D. melanogaster and C. elegans (Millward et al. 1995, Geng et al. 2000). For the catalytic domain, the deduced amino acid sequences of RsNdr1 and RsNdr2 show about 50% identity to those of the aforementioned protein kinases, though the organisms from which the protein kinases were isolated diverge at the kingdom level (Fig. 5). In addition to the sequence similarity, RsNdrs and these protein kinases share an unusual structural feature: they have an insertion of amino acids (53 and 58 residues for RsNdr1 and RsNdr2, respectively) between subdomains VII and VIII of the kinase domain. The insert is present in all of the protein kinases in this subgroup, though the functional significance of the insert is largely unknown except for animal Ndr, where the insert is used as a nuclear localization signal. A similar insert is found in some other protein kinases (e.g., plant PVPK-1 and yeast DBF2) belonging to a different subgroup of AGC protein kinases (Smith et al. 1997). The insert has not been found in any of the authentic AGC protein kinases. From these data, we consider RsNdrs to be plant orthologues of the fungal cotI-like protein kinases and animal Ndr protein kinases.

View Details

Fig. 5. A phylogenetic tree showing the relative relationship of RsNdrs to selected protein kinases in the AGC group. For the abbreviations of the names of a protein kinases, refer to the text.

Although the functions of RsNdrs have not yet been clarified, it is interesting to note that both the fungal cot-1-like protein kinases and Ndr protein kinases influence cell division and/or cell morphology through the control of cell polarity (Tamaskovic et al. 2003). Since a similar role for RsNdrs is expected in plant cells, it will be interesting to examine the phenotype of plants in which the expression of RsNdrs is altered. RsNdr-like sequences are also found in other plant species such as A. thaliana, N. tabacum, S. oleracea and M. crystallinum, though none of their protein products has been characterized. When we started the experiments reported here, only two Arabidopsis ESTs (N65509 and N64988) relevant to Ndr were deposited in the databases, but by now it has been reported that several (probably eight) genes, including AtNdr1 (DDBJ accession no. AB047278), may encode putative Ndr protein kinases in Arabidopsis (Wang et al. 2003). We conclude that RsNdr1 and RsNdr2 are members of a new protein kinase superfamily that exist universally across the plant, animal and fungal kingdoms.

Expression of RsNdrs.

To examine the expression patterns of RsNdrs, we performed Northern blot and RT-PCR analyses. Although the signals were very weak, Northern blot analysis using an RsNdr1 probe showed an interesting result: mRNAs which hybridized with the probe showed differences in molecular size among the organs. This result suggests that RsNdr1 might be regulated by a processing or splicing event. In N. crassa, a transcript of cot-1 is subjected to photoregulation, resulting in the change of a translation initiation codon (Lauter et al. 1998). Further work is needed to clarify the significance of this observation.

Protein kinase activity of RsNdr proteins expressed in E. coli.

In order to test the protein kinase activity of the products, fusion proteins of RsNdrs were expressed in E. coli. The expression vector pMAL-c2x was employed to construct a fusion plasmid, pMAL-Ndr1. A moderate amount of the fusion protein expressed using this construct was detected in the soluble fraction of total E. coli proteins.

When the protein kinase activity in the E. coli cell lysate expressing pMAL-Ndr1 was measured, activity was observed toward myelin basic protein, indicating that RsNdr1 encodes a functional protein kinase. The fact that expression of a kinase-dead version of RsNdr1 yielded no such activity supports this conclusion (Fig. 4). Purification of the fusion protein, however, led to the loss of enzymatic activity for unknown reason(s). The fusion protein may be unstable and easily denatured during the process of purification, or additional factor(s) required for the full activity of the enzyme might be present in the lysate of E. coli cells. In this assay, the protein kinase activity of RsNdr1 was not changed by adding cAMP or Ca²⁺, well-known regulatory factors for PKA and PKC, respectively, to the reaction mixture (data not shown). Purification of the protein from the lysate would make possible studies to determine the regulatory factor(s) required for the full activity of the RsNdr1. As for RsNdr2, protein kinase activity was not detected in E. coli lysate harboring a fusion plasmid, although RsNdr2, like RsNdr1, contains all 12 catalytic domains in its ORF (Fig. 2). The RsNdr2 fusion protein was largely insoluble, and this may be why the activity couldn’t be detected under our assay conditions. Several proteins smaller than the expected one were detected using anti-MBP serum in the soluble fraction of E. coli lysate with RsNdr2. This may indicate that significant degradation of the protein occurred in E. coli. Alternatively, the three substrates used here may have been inappropriate for RsNdr2 protein. In some cases, it is difficult to demonstrate protein kinase activity in bacterial expression systems because of the lack of an appropriate post-translational modification system which is, in some cases, very important for enzymatic activity. Therefore, expression systems other than E. coli, such as yeast, mammalian cells and baculovirus, should be used to test the protein kinase activity of RsNdr2. Further work is in progress to characterize RsNdrs.

We thank the late Prof. Dr. Jeff Schell and the members of the laboratory in the Max-Planck-Institut für Züchtungsforschung, Kolen, since the idea for this work evolved while T. T. was working in that laboratory as a visiting scientist. This work was supported in part by the Foundation for Bio-venture Research Center from the Ministry of Education, Culture, Sports, Science and Technology, Japan.