A Plant 126-kDa Phosphatidylinositol 4-Kinase with a Novel Repeat Structure

Phosphatidylinositol metabolism plays a central role in signaling pathways in animals and is also believed to be of importance in signal transduction in higher plants. We report here the molecular cloning of a cDNA encoding a previously unidentified 126-kDa phosphatidylinositol (PI) 4-kinase (AtPI4Kβ) from the higher plant Arabidopsis thaliana. The novel protein possesses the conserved domains present in animal and yeast PI 4-kinases, namely a lipid kinase unique domain and a catalytic domain. An additional domain, approximately 300 amino acids long, containing a high percentage (46%) of charged amino acids is specific to this plant enzyme. Recombinant AtPI4Kβ expressed in baculovirus-infected insect (Spodoptera frugiperda) cells phosphorylated phosphatidylinositol exclusively at the D4 position of the inositol ring. Recombinant protein was maximally activated by 0.6% Triton X-100 but was inhibited by adenosine with an IC50 of ∼200 μm. Wortmannin at a concentration of 10 μminhibited AtPI4Kβ activity by ∼90%. AtPI4Kβ transcript levels were similar in all tissues analyzed. Light or treatment with hormones or salts did not change AtPI4Kβ transcript levels to a great extent, indicating constitutive expression of the AtPI4Kβ gene.

Synthesis and hydrolysis of phosphoinositides play an important role in the transduction of physiological signals such as hormones, growth factors, and neurotransmitters in animal cells (1,2). Sequential phosphorylation of the D4 and D5 positions of L-␣-phosphatidyl-1-D-myo-inositol (PtdIns) 1 yields phosphatidylinositol 4-phosphate (PtdIns-4-P) and phosphatidylinositol 4,5-bisphosphate (PtdIns-4,5-P 2 ). PtdIns-4,5-P 2 is hydrolyzed to inositol 1,4,5-trisphosphate, a stimulator of calcium release from intracellular stores (1), and diacylglycerol, an activator of some protein kinase C isoforms (3), by phosphoinositide-specific phospholipase C (PI-PLC). In addition to their classical function as precursors of the second messengers diacylglycerol and inositol 1,4,5-trisphosphate, phosphoinositides, including those phosphorylated at the D3 position of the inositol ring, have been shown to regulate cytoskeleton rearrangements through the association with a variety of actin-binding proteins, including profilin, gelsolin, and villin (4,5), and can also, for example, potentiate the activation of protein kinase C (6) and PI-PLC (7,8). They also constitute major regulators of membrane-trafficking (9,10). With the exception of phosphatidylinositol 5-phosphate and phosphatidylinositol 3,4,5trisphosphate, all of the inositol phospholipids found in animals have also been identified in plants. Focusing on the role of inositol lipids in plants, independent of their much questioned role in calcium signaling, PtdIns-4-P and PtdIns-4,5-P 2 are known to affect the activity of several enzymes, including the plasma membrane H ϩ -ATPase (11) and phospholipase D (12) and can interact, in vitro at least, with cytoskeletal components (13). Profilin is believed to be involved in cytoskeleton dynamics in plant cells as it is in animal cells, through interaction with PtdIns-4,5-P 2 and actin (14). PI 4-kinase catalyzes the phosphorylation of PtdIns to Ptd-Ins-4-P, the first committed step toward the synthesis of Ptd-Ins-4,5-P 2 , and therefore represents a potentially crucial point of regulation of the phosphatidylinositol-dependent pathways. Two types of PI 4-kinases, II and III, differing in size and sensitivity to detergents and adenosine, have been identified in a wide range of tissues and cellular compartments (15,16). Type I PI kinases phosphorylate the D3 position of the inositol ring and are therefore referred to as PI 3-kinases (17).
cDNAs encoding functional PI 4-kinases have been isolated from animals and Saccharomyces cerevisiae. Putative PI 4-kinase clones have been identified in Schizosaccharomyces pombe, Dictyostelium discoideum (18), and Caenorhabditis elegans (19). The proteins encoded by these genes have now been grouped into two distinct subfamilies, 1.1 and 1.2, based on sequence and structure similarities (20). Subfamily 1.1 is represented by proteins of 68 -122 kDa and subfamily 1.2 by proteins of 200 -230 kDa (with the exception of human PI4K␣, which is a 97-kDa protein (21)). The comparison of the primary structures of these proteins has enabled the identification of several conserved domains. All PI 4-kinases possess a catalytic domain of about 230 amino acid residues, which constitute the C-terminal part of those proteins, and a so-called lipid kinase * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The  1 The abbreviations used are: PtdIns, L-␣-phosphatidyl-1-D-myo-inositol; PtdIns-4-P, PtdIns 4-phosphate; PtdIns-4,5-P 2 , PtdIns 4,5-bisphosphate; PI-PLC, phosphoinositide-specific phospholipase C; LKU, lipid kinase unique; PH, Pleckstrin homology; PCR, polymerase chain reaction; kb, kilobase pair(s); PAGE, polyacrylamide gel electrophoresis; HPLC, high performance liquid chromatography; bp, base pair(s). unique (LKU) domain of about 100 residues, the location of which varies between the different isoforms. The larger isoforms as well as PI4K␣ (21) all contain a PH domain that separates the LKU and catalytic domains in these proteins. In addition, it was recently noted that yeast PIK1 and a soluble PI 4-kinase from rat have a region in common that could not be detected in any other PI 4-kinases (22).
Studies conducted with plant tissues have demonstrated the presence of PI 4-kinase activity in plasma membranes (23), the cytosol (24), the cytoskeleton (13), and nuclei (25). The PI 4-kinase activity present in the plasma membrane of carrot cells responded to osmotic stress, cell wall-digesting enzymes, and light (26,27). After solubilization from plasma membranes, this activity could be activated by a soluble 49-kDa protein, PIK-A49, which was also able to bind and bundle actin and was identified as an elongation factor 1␣ (28). This activation was subsequently shown to be dependent on the phosphorylation status of PIK-A49 (29). Recently, a partial cDNA from Arabidopsis thaliana encoding a putative PI 4-kinase, AtPI4K␣, was isolated. The deduced amino acid sequence corresponding to this cDNA comprised domains highly similar to the LKU, PH, and catalytic domains of known PI 4-kinases and, thus, demonstrated that plants posses at least one protein structurally related to known PI 4-kinases. The PH domain present in this plant protein was expressed in bacteria and shown to bind phosphatidic acid, PtdIns-4-P, and PtdIns-4,5-P 2 (30).
We are interested in the molecular and genetic analysis of cellular processes involving phosphoinositides in higher plant cells. We describe here the molecular cloning, heterologous expression, and biochemical characterization of a PI 4-kinase, designated AtPI4K␤, from A. thaliana.

EXPERIMENTAL PROCEDURES
Material-Enzymes used for DNA restriction and modification were purchased from Boehringer Mannheim and New England Biolabs (Danvers, MA). DNA primers for polymerase chain reaction (PCR) were obtained from TibMolbiol (Berlin, Germany). [␣-32 P]dCTP and [␥-32 P]ATP were obtained from ICN (Meckenheim, Germany). PtdIns was obtained from Sigma. Unless otherwise indicated, other chemicals were purchased from Boehringer Mannheim, Merck, or Sigma.
Bacteria and Plants-Escherichia coli strain XL-1 Blue (Stratagene, Heidelberg, Germany) was used for DNA cloning procedures and for library screening. Strain DH10␣Bac (Life Technologies, Inc.) was used for in vivo construction of recombinant bacmids employed for protein expression in insect cells. A. thaliana C24 plants were grown in a phytotron in soil (GS90; Gebr. Patzer, Sinntal Jossa, Germany) with an 16-h light (22°C) and 8-h dark (15°C) period. Plants grown in liquid culture medium were obtained as follows. Seed from A. thaliana Landsberg erecta were germinated on solid AM medium (1/2 MS, 1% sucrose) for 5 days. Subsequently, plantlets were transferred to liquid MS medium (31) containing 1% (w/v) sucrose and cultured for 2 weeks under gentle shaking conditions. For Northern analysis, plants were treated for 5 days (see Fig. 6).
Isolation of the AtPI4K␤ cDNA and Sequence Analysis-DNA manipulation was performed using standard protocols, as e.g. described by Sambrook et al. (32). A 1.8-kb partial cDNA fragment with a corresponding amino acid sequence showing high similarity to mammalian PI 4-kinases was originally isolated from a potato leaf epidermal fragment cDNA library. 2 This fragment was used to screen an A. thaliana genomic library in phage EMBL 3 (kindly provided by U. Uwer, Max Planck Institute of Molecular Plant Physiology, Golm, Germany). Hybridization was performed under low stringency as described (33). Candidate phages hybridizing to the potato cDNA were used as templates to amplify putative PI 4-kinase encoding DNA fragments via PCR with degenerate forward (5Ј-AARTCNGGNGAYGAYTGYMG-3Ј) and reverse (5Ј-CCYTCNGCRTCNDARTCCAT-3Ј) primers. These primers were designed according to the partial potato sequence and a homologous rice EST (accession number D24320). A 1.2-kb amplicon was found to represent a genomic fragment homologous to PI 4-kinase-coding regions. This fragment was subsequently used to screen an A. thaliana hypocotyl ZAP II cDNA library (CD4 -16; obtained from the Arabidopsis Stock Center, Columbus, OH) under stringent conditions as described previously (33). Plaque-purified phage clones were converted to pBluescript derivatives using helper phage ExAssist according to the supplier's (Stratagene) instructions. Clone pAtPI4K␤, which contained the longest cDNA insert, was used for further analyses.
DNA sequencing was done by MWG Biotech (Munich, Germany). Computational analysis was performed with the help of the programs of the Wisconsin Genetics Computer Group (GCG Package, Version 8.1 (34)). The FASTA (35) and BLAST (36) search programs were used for sequence comparisons on DNA and amino acid sequences in Gen-Bank TM , EMBL, dbEST, and SwissProt data bases. Sequence alignments were performed using the BESTFIT program. Short repetitive sequences within the AtPI4K␤ protein were identified with the help of the SAPS program (37). 3 Protein Expression in Insect Cells-Insect (Spodoptera frugiperda Sf21) cells (Invitrogen, Leek, The Netherlands) were cultivated as monolayer cultures as described (38) at 27°C in TNM-FH medium (Sigma) supplemented with 10% fetal calf serum. Expression of recombinant AtPI4K␤ protein, fused to a polyhistidine tag at its N terminus, was achieved with the help of the Bac-to-Bac Expression System from Life Technologies, Inc. Transfer plasmid pFB-His-PI4K was constructed as follows. A DNA fragment encompassing an 820-base pairlong 5Ј fragment of the AtPI4K␤ cDNA was amplified via PCR using forward (ATCGGGATCCATGCCGATGGGACGCTTTCTA-3Ј (added BamHI site underlined; start ATG of the AtPI4K␤-coding region in italics) and reverse (5Ј-ATCGGCGGCCGCCTCTGAGTTAGGTATT-GGTTCA-3Ј (added NotI site underlined) primers and cloned into plasmid pFastBacHTb (Life Technologies, Inc.). The PCR-amplified fragment in the resulting plasmid pFBH-PI4Kn was sequenced to confirm that no nucleotide exchanges occurred during the PCR amplification. Subsequently, a 2.6-kb-long fragment of the AtPI4K␤ cDNA was inserted via PvuII and NotI restriction sites into plasmid pFBH-PI4Kn. This final cloning step reconstituted the complete AtPI4K␤-coding region. Recombinant baculoviruses obtained after transfection of insect cells with bacmid DNA (i.e. E. coli/baculovirus shuttle vector) were amplified twice to yield high titer virus stocks. For expression of recombinant protein, 9 ml of virus stock was used to infect cells (60 -80% confluency) in 150-cm 2 culture flasks.
Biochemical Analysis of Recombinant AtPI4K␤-Recombinant AtPI4K␤ protein was purified from insect cells 3 days after infection. Cells were harvested, washed once with ice-cold phosphate-buffered saline buffer (10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 0.14 M NaCl, 2.7 mM KCl, pH 7.3), and lysed by sonification in lysis buffer containing 50 mM Tris-HCl, pH 7.6, 2 mM EDTA, 10% glycerol, 20 g/ml aprotinin, and 50 g/ml chymostatin. Undisrupted cells were removed by centrifugation (14.000 rpm for 10 min, Centrifuge 5417, Eppendorf), and the supernatant was analyzed by SDS-PAGE or used for protein purification. Recombinant protein was batch-purified under native conditions using nickel nitrilotriacetic acid resin according to manufacturer's (Qiagen) instructions. Yield of recombinant AtPI4K␤ from a representative purification (24 ml of insect cell culture) was 0.2 mg of protein. Purity of recombinant AtPI4K␤ was above 95% as estimated from Coomassiestained SDS-PAGE protein gels. Protein concentrations were determined by the method of Bradford (1976) using bovine serum albumin as standard (41). PI 4-kinase activity was determined at 37°C in 50 l of reaction buffer (50 mM Tris-HCl, pH 7.4, 80 mM KCl, 10 mM MgCl 2 , 2 mM EDTA, 5 g PtdIns (Sigma), 0.2% Triton X-100 (unless otherwise indicated)), and 0.1 g of affinity-purified protein. The reaction was started by the addition of 50 nM [␥-32 P]ATP (Ͼ600 Ci/mmol; ICN). The reaction was stopped after 30 min by adding 0.8 ml of cold chloroform/methanol/H 2 O (1/2/0.8; v/v/v). To extract lipids, 0.4 ml of chloroform/2.4 N HCl (1/1; v/v) was added, and the reaction mixture was vortexed and spinned for 30 s at 12,000 rpm. The upper phase was removed, and an equal volume of chloroform/methanol/HCl (5/245/235; v/v/v) was used to wash the chloroform phase a second time. An aliquot of the chloroform phase (15 l) was applied to thin-layer chromatography (TLC) plates (Silica gel 60 plates; Merck). Chromatograms were developed with a chloroform/ methanol/concentrated NH 3 /H 2 O (45/35/2/8, v/v/v/v) solvent mixture. Plates were exposed to Kodak X-Omat AR films. TLC revealed that 32 P in the chloroform phase was incorporated into PtdIns-P only. To determine the effect of Triton X-100, adenosine, and wortmannin (Calbiochem, Bad Soden, Germany) on PI 4-kinase activity, liquid scintillation counting was performed on the dried washed chloroform phase.
To determine the effect of wortmannin on PI 4-kinase activity, 1 l of stock solution (10 M to 1 mM, wortmannin dissolved in ethanol) was added to the reaction mixture. Ethanol was used in control experiments. All experiments were performed at least three times on independent preparations of purified protein. The data presented are the mean ϮS.D.
HPLC Analysis of the Reaction Product-The products of enzyme assay were deacylated, mixed with 3 H standards, and resolved by Partisphere SAX HPLC according to Brearley et al. (42).
RNA Blot Analysis-Material for RNA blot analysis of steady state mRNA levels was harvested from well watered A. thaliana plants (Fig.  6, top left) or from plants cultivated in liquid medium (Fig. 6, top right and bottom). RNA was prepared according to the protocol of Logemann et al. (43). Total RNA was quantitated spectrophotometrically at 260 nm. RNA (30 g/lane) was electrophoretically separated in denaturing 15% (v/v) formaldehyde, 1.5% (w/v) agarose gels and blotted onto Hybond-N ϩ membranes (Amersham Pharmacia Biotech). RNA was fixed to the membrane via UV-cross-linking (Stratalinker, Stratagene). An 0.8-kb-long 5Ј fragment of the AtPI4K␤ cDNA was used as [␣-32 P]dCTPlabeled hybridization probe. Membranes were hybridized at 65°C in 250 mM sodium phosphate buffer, pH 7.2, containing 7% SDS, 1% bovine serum albumin, and 1 mM EDTA. Washes were performed at 65°C in 3ϫ SSC (1ϫ SSCϭ 0.15 M NaCl and 0.015 M sodium citrate), 0.5% SDS for 15 min and in 0.2ϫ SSC, 0.5% SDS for 20 min. Blots were exposed to Kodak X-Omat AR films between intensifying screens for 2-3 days at Ϫ70°C.

RESULTS
Cloning of AtPI4K␤-A partial cDNA from potato encoding a polypeptide with high similarity to animal and yeast PI 4-kinases was used to isolate via various intermediate cloning steps (for details see "Experimental Procedures") a corresponding cDNA from A. thaliana. The longest hybridizing cDNA (4349 nucleotides) obtained from this library, designated AtPI4K␤ to differentiate it from the partial clone (AtPI4K␣) isolated by Stevenson et al. (30), contained an open reading frame of 3366 nucleotides, representing a 1121-amino acid polypeptide with a calculated molecular mass of 126 kDa. The putative initiation codon was preceded by an in-frame stop codon; therefore, this clone represented a full-length cDNA (Fig. 1A). The calculated isoelectric point of 5.5 of AtPI4K␤ is similar to the isoelectric point of 5.8 determined for the partially purified PI 4-kinase from Catharanthus roseus (44). A genomic fragment harboring the entire AtPI4K␤-coding region was present in a P1 clone (MHJ24; GenBank TM accession number AB008266) sequenced by the Kazusa DNA Research Institute. A comparison between the two DNA sequences revealed the presence of 16 exons (varying in size between 49 and 2163 base pairs) and 15 introns (sizes between 95 and 511 base pairs) in the genomic fragment (Fig. 1B). The computer-predicted exon/intron structure was mostly correct with the exception of the 3Ј region of the gene, where three instead of two exons were found (marked in Fig. 1B); an additional exon was detected at the 3Ј end of the gene (Fig. 1B). The complete AtPI4K␤ gene had a size of approximately 8 kb, including 5Јand 3Ј-untranslated regions and putative 5Ј regulatory elements. Mapping analysis indicated that the AtPI4K␤ gene is located at the bottom of chromosome V, between markers CIC9B5L and T04492.

Structural Organization of the A. thaliana PI 4-Kinase Protein-
The deduced AtPI4K␤ protein sequence possessed two regions highly similar to the LKU and catalytic domains of known PI 4-kinases ( Fig. 2A). The LKU domain in AtPI4K␤ was located close to the N terminus and was 20 -37% identical to LKU domains from previously identified PI 4-kinases (Fig.  2B). The C terminus of AtPI4K␤ contained the putative catalytic domain, which was 43-57% identical to the corresponding domains from other PI 4-kinases (Fig. 2C). In addition, the A. thaliana protein described here, like all the members of subfamily 1.1, lacks a PH domain, a domain that characterizes members of subfamily 1.2. These observations indicated that AtPI4K␤ represented a new member of the subfamily 1.1 of PI 4-kinases, with a structure resembling most closely that of the yeast PIK1 protein. Overall, AtPI4K␤ was more similar to subfamily 1.1 PI 4-kinases than to the other A. thaliana partial PI 4-kinase, AtPI4K␣, and was most closely related to a putative PI 4-kinase from D. discoideum, with 30% identical and 52% similar amino acid residues.
Further sequence analyses showed that AtPI4K␤ also possessed a region similar to part of a sequence previously thought to be shared only by yeast PIK1 and a soluble 92-kDa PI 4-kinase from rat brain (Fig. 2D). The region common to these three proteins, designated NH, is approximately 90 amino acid residues in length.
Hydropathy analysis identified a relatively hydrophilic segment within the N-terminal half of AtPI4K␤, whereas the rest of the protein had a more amphipathic character. Part of the hydrophilic segment, covering amino acids 239 -560, contained a high percentage (46%) of charged amino acid residues, with 25% acidic (glutamate and aspartate) and 21% basic (lysine, arginine, and histidine) amino acids. This domain was not present in any of the previously identified PI 4-kinases. A more careful analysis of this region revealed the presence of a repeated motif. Fig. 3 shows that the repeated sequence, which occurs 11 times between amino acid position 212 and 508 (see Fig. 1A), has a length of 19 -20 amino acids with a highly conserved inner core sequence of hydrophobic/positively charged/hydrophobic amino acids. This core sequence was Nterminal-flanked by acidic amino acids (aspartate or glutamate), and C-terminal-flanked by alternating positively and negatively charged amino acids.
Biochemical Characterization of Recombinant AtPI4K␤-To investigate whether AtPI4K␤ encodes a functional PI 4-kinase, we expressed it with a polyhistidine tag at its N terminus in baculovirus-infected insect (S. frugiperda) cells. As shown in Fig. 4A, affinity-purified recombinant AtPI4K␤ appeared as a single band on a protein gel (Fig. 4A, left panel, lane 4). The purity of the recombinant protein was estimated to be higher than 95%, because almost no additional bands were present on the gel. Although it was not possible to detect recombinant AtPI4K␤ in the corresponding crude extract after Coomassie staining, an antibody directed against the N-terminal polyhistidine tag showed that recombinant protein was present in the crude extract (Fig. 4A, right panel, lane 2). No cross-reacting protein was detected in cells expressing an unrelated protein without a polyhistidine tag. The molecular mass of the recombinant AtPI4K␤ protein was estimated to be 125-130 kDa, which closely agrees with the calculated mass of 126 kDa.
When analyzed in vitro, recombinant AtPI4K␤ yielded a phospholipid that on TLC, co-chromatographed with authentic PtdIns-4-P (Fig. 4B). The reaction product was deacylated and analyzed by HPLC on a Partisphere SAX column. The deacylated product co-eluted exactly with [ 3 H]glycerophosphoinositol 4-phosphate after glycerophosphoinositol 3-phosphate (Fig.  4C). We also analyzed the levels of PtdIns-4-P in insect cells. Cells expressing the AtPI4K␤ consistently showed 25-30% in-crease in PtdIns-4-P levels as compared with mock-infected cells (data not shown). These results confirm that AtPI4K␤ is a functional PI 4-kinase.
Triton X-100 activates animal PI 4-kinases (15) and is also able to activate the partially purified enzymes from carrot (45) and C. roseus suspension culture cells (44). The activity of recombinant AtPI4K␤ was almost unaffected by Triton X-100 concentrations below 0.3% (w/v) but increased 4-to 5-fold when the detergent concentration was increased to 0.6% (Fig. 5A).
Adenosine inhibits mammalian PI 4-kinases previously classified as type II more than it does type III enzymes (20). When tested on AtPI4K␤, a moderate inhibitory effect was observed, i.e. 50% inhibition was reached at 200 M adenosine (Fig. 5B). Maximal activity of recombinant AtPI4K␤ protein was approximately 12 nmol/min/mg of protein (in the presence of 0.6% Triton X-100).  Fig. 3). B, exon/intron structure of the AtPI4K␤ gene. Exons are presented as boxes. The four marked exons in the 3Ј region of the gene were identified via comparison with the AtPI4K␤ cDNA and were found to differ from the original computer prediction (see accession number AB008266). Numbers indicate exon sizes (nucleotides).
Wortmannin, a hydrophobic steroid-related compound from the fungus Talaromyces wortmannii, is a potent inhibitor of mammalian PI 3-kinases (46). Although PI 4-kinases were not first believed to be inhibited by wortmannin, it has recently been demonstrated that the yeast PI 4-kinase encoded by the STT4 gene is inhibited by approximately 95% at 10 nM wortmannin (47). Wortmannin-sensitive PI 4-kinases have also been cloned from human (48) and bovine (49), although sensitivity of the PI 4-kinase is at least 10 times lower than that of P13-kinase. As is shown in Fig. 5C, wortmannin had little effect at concentrations up to 0.1 M but inhibited the plant recombinant enzyme by approximately 90% at 10 M.
Northern Blot Analyses Indicate Almost Constitutive Expression of the AtPI4K␤ Gene-The full-length AtPI4K␤ cDNA was used as hybridization probe to study transcript levels. A single transcript of approximately 4.5 kb was detected in leaves, roots, flowers, and stems of A. thaliana (Fig. 6, top left). AtPI4K␤ transcript levels were similar in the various tissues analyzed. An almost identical result was obtained using quantitative reverse transcription PCR (data not shown). These experiments also indicated the presence of AtPI4K␤ transcript in epidermal fragments, a tissue preparation highly enriched for stomatal guard cells (data not shown). AtPI4K␤ transcript levels in leaves were closely similar in the light and dark (Fig.  6, top right). Treatment of small plantlets with hormones, CaCl 2 , or NaCl had no effect on AtPI4K␤ mRNA levels (Fig. 6,  bottom). The AtPI4K␤ gene appeared therefore to be constitutively expressed in A. thaliana.

DISCUSSION
Although PI 4-kinase activity has long been detected, characterized, and partially purified from higher plants, cloning and functional expression of a plant PI 4-kinase has not yet been reported. Here we describe the cloning of a cDNA from A. thaliana that encodes a functional 126-kDa PI 4-kinase. The size of this protein is similar to the size of the smaller of the PI 4-kinases isolated from yeast, D. discoideum, and animals, which is in the range of 90 -122 kDa. Recently, Okpodu et al. (45) partially purified a soluble PI 4-kinase from carrot suspension culture cells. The molecular mass of this protein was estimated to be 83 kDa. Two distinct PI 4-kinase activities have recently been partially purified from spinach plasma membranes, with estimated sizes of 65 and 120 kDa. 4 No peptide sequences for any of the partially purified plant PI 4-kinases are available. Therefore it is not possible to know whether any of the partially purified plant PI 4-kinases corresponds to homologs of AtPI4K␤ or AtPI4K␣.
Sequence analyses showed that AtPI4K␤ possesses the two conserved domains present in all PI 4-kinases, namely the LKU and the kinase catalytic domains, but lacked a PH domain. These structural features indicate that AtPI4K␤ is a new member of subfamily 1.1 of PI 4-kinases. The molecular structure of AtPI4K␤ most resembles that of the yeast PIK1 and the Dictyostelium DdPI4K proteins. We also identified a novel domain, NH, that corresponds to part of the sequence that was previously identified only in the yeast PIK1 protein and a rat soluble PI 4-kinase (22). Surprisingly, AtPI4K␤ is the only known PI 4-kinase to possess a unique repetitive motif constituted of 11 repeats of a charged core unit. Interestingly, three putative PtdIns-P kinases from Arabidopsis (Refs. 50 and 51 and accession number U95973) also contain a repeated motif, whereas none of the animal enzymes possess such a repeated motif. No clear function can be assigned to the LKU, the NH, or the domain with the repetitive motif; they may play a role in the interaction of PI 4-kinases with other proteins and/or membrane structures.
The partial PI 4-kinase isolated from A. thaliana, AtPI4K␣ (30), possesses a PH domain, and its structure resembles that of members of subfamily 1.2 of PI 4-kinases. It was shown that the PH domain of AtPI4K␣ can bind PtdIns-4-P and PtdIns-4,5-P 2 but not PtdIns. It was suggested that AtPI4K␣, by binding phosphoinositides at its PH domain, could be involved in the regulation of actin polymerization. It appears, thus, likely that the PH domain of type 1.2 PI 4-kinases is not responsible for binding their substrate. Consequently, it is possible that substrate binding in PI 4-kinases is controlled by the LKU domain or a domain conserved structurally but not at the sequence level.
The catalytic properties of the recombinant plant protein, i.e. its stimulation by low concentrations of Triton X-100 as well as its moderate inhibition by adenosine, although they do not match exactly, still resemble those of the PI 4-kinases previ- The lipid products obtained as in B were purified, deacylated, and analyzed by HPLC as described (42). The migration of 3 H standards glycerophosphoinositol 3-phosphate (GroPIns3P) and glycerophosphoinositol 4-phosphate (GroPIns4P) is indicated with open symbols; 32 P product is indicated with closed symbols.
ously classified as type III more than those classified as type II. As recently observed with other PI 4-kinases, the AtPI4K␤ is inhibited by wortmannin at concentrations that are significantly higher than those required to inhibit PI 3-kinases. The residue corresponding to the Lys residue of the PI 3-kinase PI3K␣ at which wortmannin binds covalently (52) is conserved in AtPI4K␤ as it is in all other cloned PI 4-kinases, including AtPI4K␣. Recently, wortmannin at high concentrations was shown to inhibit PI 3-kinase and PI 4-kinase activities in tobacco cells in vivo as well as protein sorting to the plant vacuole (53). These results suggest that the sorting of proteins to the plant vacuole is controlled by at least two different mechanisms, one of which is wortmannin-sensitive and may involve phosphoinositides (53). It has been shown in animal cells that a wortmannin-sensitive PI 4-kinase enzyme was in-volved in the formation of agonist-stimulated inositol trisphosphate (54). Because plants appear to express similar PI 4-kinase isoforms as animals, it will be of interest to test whether any of the factors that affect phosphoinositide levels in plants, such as light, mastoparan, and fungal elicitors (55), involve the response of a wortmannin-sensitive PI 4-kinase enzyme.
The sequences available so far indicate that in animals each of the two types of PI 4-kinases is represented by several splice variants of one single gene. The whole genome of S. cerevisiae contains only two PI 4-kinase genes. It is now clear that plants also contain subfamily 1.1 and subfamily 1.2 PI 4-kinases. Because PI 4-kinase activities have been detected in most cellular compartments, it is probable that the domains identified in PI 4-kinases are involved in the targeting of PI 4-kinase isoforms to various cellular compartments through interactions with other proteins or lipids. The similarity in structure of the PI 4-kinases found in the different phyla contrasts markedly with the current data available for PI-PLC. Three to four genes have been isolated from A. thaliana (56 -58), potato (59), and soybean (60). Based on protein structure and biochemical properties, all plant PI-PLC isoforms appear to belong to a single family, which is most closely related to the ␦-type of mammalian PI-PLCs, but all lack a PH domain. Because phosphoinositides are now known to be important in processes not involving the canonical second messengers inositol trisphosphate and diacylglycerol, it will be of great interest to compare the regulation and function of PI 4-kinases between animal and plants. In addition, whereas the plant PI-PLC isoforms are expressed in tissue-and/or environmental-specific manner (56,59), the plant PI 4-kinase reported here is ubiquitously expressed in the plant. Future experiments will tell whether plants possess additional subfamily 1.1 PI 4-kinase genes whose expression patterns are developmentally and/or environmentally regulated.