A Role for Rat Inositol Polyphosphate Kinases rIPK2 and rIPK1 in Inositol Pentakisphosphate and Inositol Hexakisphosphate Production in Rat-1 Cells*

Over 30 inositol polyphosphates are known to exist in mammalian cells; however, the majority of them have uncharacterized functions. In this study we investigated the molecular basis of synthesis of highly phosphorylated inositol polyphosphates (such as inositol tetrakisphosphate, inositol pentakisphosphate (IP5), and inositol hexakisphosphate (IP6)) in rat cells. We report that heterologous expression of rat inositol polyphosphate kinases rIPK2, a dual specificity inositol trisphosphate/inositol tetrakisphosphate kinase, and rIPK1, an IP5 2-kinase, were sufficient to recapitulate IP6 synthesis from inositol 1,4,5-trisphosphate in mutant yeast cells. Overexpression of rIPK2 in Rat-1 cells increased inositol 1,3,4,5,6-pentakisphosphate (I(1,3,4,5,6)P5) levels about 2–3-fold compared with control. Likewise in Rat-1 cells, overexpression of rIPK1 was capable of completely converting I(1,3,4,5,6)P5 to IP6. Simultaneous overexpression of both rIPK2 and rIPK1 in Rat-1 cells increased both IP5 and IP6 levels. To reduce IPK2 activity in Rat-1 cells, we introduced vector-based short interference RNA against rIPK2. Cells harboring the short interference RNA had a 90% reduction of mRNA levels and a 75% decrease of I(1,3,4,5,6)P5. These data confirm the involvement of IPK2 and IPK1 in the conversion of inositol 1,4,5-trisphosphate to IP6 in rat cells. Furthermore these data suggest that rIPK2 and rIPK1 act as key determining steps in production of IP5 and IP6, respectively. The ability to modulate the intracellular inositol polyphosphate levels by altering IPK2 and IPK1 expression in rat cells will provide powerful tools to study the roles of I(1,3,4,5,6)P5 and IP6 in cell signaling.

Several recent works suggest possible functions of highly phosphorylated inositol polyphosphates and/or their kinases. For example, IP 5 has been shown to modulate human immunodeficiency virus, type 1, Gag protein assembly (14), and IP 6 has been reported to bind the clathrin assembly proteins AP2 and AP3 thus inhibiting clathrin cage assembly in vitro (15)(16)(17)(18). Additionally IP 6 stimulates non-homologous DNA end joining of double strand DNA breaks by binding to the Ku70/80 subunits of DNA-dependent protein kinase in vitro (19 -21). Yeast genetic data also suggest that these kinases and/or their inositol polyphosphate products are involved in transcription of specific genes, mRNA export, and chromatin remodeling (5)(6)(7)(22)(23)(24). However, most of the data mentioned above were determined in vitro or in yeast, and evidence in mammalian cells is still lacking. Additionally some of the data obtained from in vitro studies may be experimental artifacts due to the effect of the highly polarized negative charge of these inositol polyphosphate molecules (25).
To address the synthetic pathway and the function of inositol polyphosphates in mammalian cells, it is important to establish a model system whereby the inositol polyphosphate levels can be modulated in vivo. In this study, we analyzed the effects of either overexpression or knock-down of inositol polyphosphate kinase(s) on inositol polyphosphate levels in Rat-1 cells. We showed the accumulation of I(1,3,4,5,6)P 5 and IP 6 by overexpression of IPK2 and IPK1. We also succeeded in decreasing cellular I(1,3,4,5,6)P 5 levels by RNA interference (RNAi). These data reveal the involvement of IPK2 and IPK1 in I(1,3,4,5,6)P 5 and IP 6 synthesis in vivo. Furthermore the ability to modulate the intracellular inositol polyphosphate levels shown here by altering IPK2 and IPK1 expression in rat cells will provide powerful tools to study the roles of I(1,3,4,5,6)P 5 and IP 6 in eukaryotic cell signaling.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Phoenix and Rat-1 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere of 5% CO 2 . Culture medium was supplemented with 100 units/ml penicillin G and 100 g/ml streptomycin. The cell lines expressing tetracycline-inducible constructs were maintained in the same condition except using 10% Tet system-approved FBS (Clontech). Retroviruses were produced by transient transfection of the plasmid into phoenix cells by the calcium phosphate method as described elsewhere (26). Rat-1 cells were infected with retrovirus-containing media in the presence of 8 g/ml hexadimethrine bromide (Sigma) for overnight. Thirty-six hours later, cells were selected in the presence of appropriate antibiotics for at least 2 weeks and then used for experiments. The concentration of antibiotics used for selection were: Geneticin, 800 g/ml; hygromycin, 400 g/ml; and puromycin, 1.5 g/ml, respectively.
RNA Preparation and Cloning of Rat IPK2 and Rat IPK1 by RT-PCR-Total RNA was isolated according to the manufacturer's recommendations using the RNeasy minikit (Qiagen). Briefly 1 ϫ 10 7 Rat-1 cells were lysed into Buffer RLT with 1% 2-mercaptoethanol and then applied to the RNeasy minicolumn. After washing the column, total RNA was eluted with 40 l of RNase-free water. RT-PCR was performed using the RobusT I RT-PCR kit (Finnzymes). Reverse transcription of 1 g of RNA was performed at 42°C for 1 h in 50 l of reaction mixture containing 1ϫ RobusT I reaction buffer, 1.5 mM MgCl 2 , 40 units of RNase inhibitor, 0.8 mM dNTPs, 0.2 M oligo(dT), 0.2 M sense and antisense primers, 5 units of avian myeloblastosis virus reverse transcriptase, and 1 unit of Dy-NAzyme EXT DNA polymerase. The subsequent PCR was done at 94°C for an initial 2 min followed by 25 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min and final extension at 72°C for 10 min. The primers were designed to amplify the entire open reading frame of the rat ortholog of IPK2 (rIPK2) (GenBank TM accession number AY014898) (10). The primers for the rat ortholog of IPK1 (rIPK1) were designed based on the mouse IPK1 ortholog (GenBank TM accession number XM_283126). The primers used for rIPK2 were: sense, 5Ј-TTT TTG TCG ACC ACC ATG GCC GCC GAG CCC CCA-3Ј (the SalI site is underlined); and antisense, 5Ј-TTT TTG ATA TCG ATT CAA CTG TCC AAG ATA CTC CG-3Ј (the ClaI site is underlined). The primers used for rIPK1 were: sense, 5Ј-AAC TCG AGC ACC ACC ATG GAA GAG GGG AAA AT-3Ј (the XhoI site is underlined); and antisense, 5Ј-AAT CTA GAA AGC TTT TAG ACC TTA TGG AGA ACT AAT GTG CCC G-3Ј (the HindIII site is underlined). The RT-PCR products were cloned into pCR2.1 TA cloning vector (Invitrogen). The substitution of Asp-127 to Ala in rIPK2 (rIPK2 D127A) was performed by PCR with the following mutagenic primers (the D127A mutation is underlined): sense, 5Ј-AAG CCC TGT ATA ATG GCC GTG AAG ATT GGG CGG-3Ј; and antisense, 5Ј-CCG CCC AAT CTT CAC GGC CAT TAT ACA GGG CTT-3Ј. All constructs were sequenced by the ABI Prism Big Dye Terminator Ready reaction kit (Applied Biosystems).
Expression and Purification of Recombinant GST Fusion Protein-The expression and purification of GST-rIPK2 and GST-rIPK1 were performed as described previously (29) with minor modifications. Briefly competent BL21 Escherichia coli cells were transformed by pGST/rIPK2 or pGST/rIPK1. One liter of LB was cultured at 37°C to an absorbance of 0.8 -1.0 at 600 nm. Expression was induced by the addition of 0.1 mM isopropyl ␤-D-thiogalactopyranoside (final concentration). The cells were grown for 20 h at 20°C, harvested by centrifugation at 4°C, and resuspended in 20 ml of ice-cold phosphate-buffered saline containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride and one Complete TM Mini protease inhibitor mixture tablet (Roche Applied Science)/10 ml of buffer). The cells were lysed by five passages through a cell cracker (a high shear fluid-processing system for cell rupture, Microfluidics Corp.). Triton X-100 was added to lysates at final concentrations of 1%, and then lysates were subjected to centrifugation at 15,000 ϫ g for 15 min at 4°C. The supernatant was then incubated with 1 ml of 50% glutathione-Sepharose slurry (Amersham Biosciences) for 1 h at 4°C with shaking. The Sepharose was washed three times with 15 ml of ice-cold phosphate-buffered saline, and the GST fusion proteins were eluted from the Sepharose with 1 ml of 10 mM reduced glutathione in 50 mM Tris-HCl (pH 8.0) and stored at Ϫ80°C in aliquots.
Cell Lysate Preparation and Western Blotting Analysis-2 ϫ 10 5 Rat-1 cells stably expressing the indicated plasmid were seeded in 60-mm culture dishes containing 5 ml of the appropriate media with or without 2 g/ml doxycycline and cultured for 2-3 days. Cells were washed twice with 5 ml of ice-cold phosphate-buffered saline and then extracted with 200 l of ice-cold 1% Nonidet P-40 TNE buffer (10 mM Tris-HCl (pH 7.6), 1% Nonidet P-40, 150 mM NaCl, 2 mM phenylmethylsulfonyl fluoride, and one Complete TM Mini protease inhibitor mixture tablet/10 ml of buffer). After a 15-min incubation on ice, the lysate was recovered by centrifugation (14,000 rpm for 15 min at 4°C). The supernatants were subjected to protein assay using bovine serum albumin as standard. One microgram of cell lysates was separated by SDS-PAGE. Immunostaining of gels was carried out using an ECL Western blotting detection system (Amersham Biosciences) using a rabbit polyclonal anti-GFP antibody (Clontech).
Fluorescence Microscopy-Rat-1 cells stably expressing GFP fusion constructs were washed with prewarmed Hanks' balanced salt solution and observed under the Olympus IX70 inverted microscope equipped with a confocal head using a LCPlan Fl 40ϫ/0.60 numerical aperture lens. Images were recorded using UltraVIEW TM imaging software (PerkinElmer Life Sciences). 1 ϫ 10 5 Rat-1 cells stably expressing the indicated plasmid were seeded into a 60-mm culture dish containing 3 ml of appropriate media. One day after incubation, cells were washed twice with 5 ml of prewarmed Medium-199 and then labeled with 10 -20 Ci/ml myo-[2-3 H]inositol for 2 days in 2 ml of Medium-199 supplemented with dialyzed 10% FBS with or without 2 g/ml doxycycline. Cells were washed twice with ice-cold phosphate-buffered saline and then harvested in 500 l of 1 N HCl. The soluble fraction containing the inositol polyphosphates was isolated as mentioned above. To determine whether the IP 5 product in the rIPK2-expressing Rat-1 cells was I(1,3,4,5,6)P 5 , the lysate was dried using a SpeedVac and resuspended to 50 mM Hepes-NaOH (pH 7.5). The lysate was incubated with 100 ng of purified recombinant GST-AtIpk1 for 30 min at 37°C.
Northern Blotting Analysis-Northern blot analysis was performed as described by Stevenson et al. (33). The entire open reading frame of each gene was used as probe and labeled using Ready-To-Go TM DNA labeling beads (Amersham Biosciences) with [␣-32 P]dCTP. Hybridization was carried out using ExpressHyb hybridization solution (Clontech) following the manufacturer's recommendations.
Colony Formation Assay in Soft Agar-5 ϫ 10 4 Rat-1 cells stably expressing the indicated plasmid were seeded into a 35-mm culture dish containing 1 ml of culture media with 0.4% agar. The dishes had been coated with 2 ml of culture media containing 0.7% agar. Cells were fed with 0.5 ml of culture media containing 0.4% agar once a week. Colony formation was scored after 3 weeks at 37°C in a humidified atmosphere of 5% CO 2 .

Cloning of Rat IPK2 and Rat IPK1
Orthologs-To investigate the function of inositol polyphosphates in mammalian cells, we first cloned rat orthologs of IPK2 and IPK1. Saiardi et al. (10) have published the cloning of rIPK2, so we used this sequence information for cloning (see "Experimental Procedures").
When we started this project, the DNA sequence information of rIPK1 was not available. However, we could use both the human and mouse sequences of IPK1. These two orthologs possess about 88% identity at the DNA level. Based on this information, we designed cloning primers for rIPK1 and performed RT-PCR (see "Experimental Procedures"). The rat ortholog of IPK1 possesses 91.9 and 96.3% of amino acid identity to the human and mouse orthologs, respectively. The rIPK1 contains the conserved motifs EXKPK, CRXC, (F/Y)CPLDL, and D(L/V)DLK(P/S)X(E/M) (13), the function of which are still unknown.
Recently the predicted open reading frame of rIPK1 was published in GenBank TM (GenBank TM accession number XM_225201). This prediction was based on GenomeScan analysis of the rat genomic sequence (GenBank TM accession number NW_047490). Our cDNA differed somewhat from the predicted intron/exon splicing identified by the GenomeScan software. Therefore, we submitted a new sequence, and Gen-Bank TM issued a new accession number (GenBank TM accession number AY823319) for rIPK1.
Complementation of Inositol Polyphosphate Production in Mutant Yeast Cells-We tested whether rIPK2 could rescue inositol polyphosphate synthesis in yeast cells lacking endogenous IPK2. In this experiment, the wild type, ipk2 null (ipk2⌬), and ipk2⌬ strain expressing either rIPK2 or rIPK2 D127A mutant were metabolically labeled with [ 3 H]inositol, and then the soluble inositol polyphosphates were analyzed by HPLC. In the wild type strain, IP 6 was the major inositol polyphosphate ( Fig. 2A). In the ipk2⌬ strain, IP 3 accumulated because Ipk2 is the major kinase responsible for the conversion of IP 3 to produce higher inositol polyphosphates such as IP 5 and IP 6 ( Fig.  2B) (6, 7). The ipk2⌬ strain harboring the rIPK2 expression plasmid complemented IP 5 and IP 6 production (Fig. 2C). These data confirmed that rIPK2 is active in yeast and is capable of converting I(1,4,5)P 3 to I(1,3,4,5,6)P 5 . Expression of the rIPK2 D127A mutant partially rescued IP 6 synthesis (Fig. 2D), consistent with our biochemical studies indicating that this mutation in the rat enzyme still possesses weak activity.
Expression and Intracellular Localization of Inositol Polyphosphate Kinase(s) in Rat-1 Cells-After confirmation of the enzymatic activity of rIPK2 and rIPK1 in vitro and in budding yeast, we expressed these genes in rat cells. An inducible mammalian expression system (Tet-On TM gene expression system, Clontech) was chosen thereby enabling the control of protein expression by the tetracycline or tetracycline analogue doxycycline (34). We expressed control GFP, yeast or rat IPK2, and yeast or rat IPK1 as GFP fusion proteins under the control of doxycycline in the Rat-1 rat embryonic fibroblast cells. Appropriate protein expression was confirmed using an anti-GFP immunoblot (data not shown). In the stable cell lines used in the absence of doxycycline, there was some leaky expression of protein, but 2 g/ml doxycycline induced expression levels of all of the fusion proteins (data not shown).
We investigated the intracellular localization of these proteins. The expression of the GFP fusion proteins was induced by 2 g/ml doxycycline, and living cells were observed under confocal microscopy. In contrast to the ubiquitous localization of GFP alone (Fig. 3A), GFP-rIPK2 was localized in the nucleus (Fig. 3C). To our knowledge, there is no other report describing rIPK2 localization, but it has been reported that the human IPK2 localizes to the nucleus, and the nuclear localization signal sequence was identified (11). This sequence is conserved between human and rat IPK2, so this nuclear localization signal sequence is likely also functional in rIPK2. It was reported that yeast Ipk2 is localized to the nucleus in yeast cells (6), but when we expressed yeast Ipk2 in Rat-1 cells, it localized to the cytosol (Fig. 3B). This difference probably reflects the difference of nuclear localization mechanisms between yeast and Rat-1 cells. Yeast Ipk1 is localized to the nuclear envelope in yeast cells (5), but both yeast Ipk1 and rat IPK1 were localized ubiquitously in Rat-1 cells (Fig. 3, D and E).
Inositol Polyphosphate Profiles in IPK2-or IPK1-overexpressing Rat-1 Cells-To address whether the overexpression of these genes effects inositol polyphosphate production, we labeled these cells using [ 3 H]inositol and then analyzed the sol-uble inositol polyphosphates by HPLC. In the untransfected (Fig. 4A) or GFP-expressing Rat-1 cells (Fig. 4B), IP 5 was the major inositol polyphosphate, and it was about 16 -18% of the total inositol polyphosphates. IP 6 was about 7-8% of total inositol polyphosphates in these cells. Both the yeast (Fig. 4C) and rat IPK2 overexpression (Fig. 4E) increased IP 5 levels about 2-3-fold compared with control cells. Overexpression of Scipk2 kin Ϫ (Fig. 4D) did not show any increase of IP 5 levels, but the rIPK2 D127A mutant (Fig. 4F) expressing Rat-1 cells showed a subtle but consistent increase in IP 5 content compared with control cells. Our data indicate that increased IPK2 activity, nuclear in the case of rIPK2 or cytoplasmic in the case of scIpk2, is sufficient to elevate IP 5 levels.
To confirm that the IP 5 isomer that accumulates in the IPK2 overexpressing Rat-1 cells is indeed I(1,3,4,5,6)P 5 , we prepared soluble radiolabeled inositol polyphosphate extracts from these cell lines and subjected the IPs to enzyme analysis. As mentioned above, AtIpk1 can phosphorylate I(1,3,4,5,6)P 5 to IP 6 but not other IP 5 species (data not shown). The IP 5 peak in the extract of cells expressing either GFP control or GFP-rIPK2 was completely phosphorylated to IP 6 by AtIpk1 in vitro (data not shown). These data show that rIPK2 synthesizes I(1,3,4,5,6)P 5 in Rat-1 cells and also shows that the major IP 5 isomer in the Rat-1 cells is I(1,3,4,5,6)P 5 , which is consistent with previous reports (35).
Next we investigated the effect of IPK1 overexpression on inositol polyphosphate production in the Rat-1 cells. Both the yeast Ipk1 and rat IPK1 overexpression increased the IP 6 level to about 20% of total inositol polyphosphates (Fig. 4, G and H). This was about a 2.5-fold elevation of IP 6 levels compared with control cells. Interestingly the I(1,3,4,5,6)P 5 levels were de- creased to undetectable levels in these cells, indicating that levels of endogenous rIPK1 are highly regulated and/or rate determining.
Since all of the I(1,3,4,5,6)P 5 was converted to IP 6 when the cells were overexpressing IPK1 and since I(1,3,4,5,6)P 5 was increased with IPK2 overexpression, we attempted to generate high levels of IP 6 in Rat-1 cells by overexpressing IPK2 and IPK1 simultaneously. Fig. 5, A and C, shows the HPLC profiles from Rat-1 cells expressing only Myc-tagged rIPK1 under the control of the tetracycline-inducible system. In the absence of doxycycline, the HPLC profile was similar to control cells (Fig.  5A). The presence of doxycycline induced IPK1 expression, and the I(1,3,4,5,6)P 5 peak shifted to IP 6 (Fig. 5C). Fig. 5, B and D, shows the HPLC profiles from Rat-1 cells that co-expressed GFP-tagged rIPK2 with Myc-tagged rIPK1. Because the expression of GFP-rIPK2 in these cells was not under the control of the tetracycline-inducible system, I(1,3,4,5,6)P 5 was increased in the absence of doxycycline (Fig. 5B). With the doxycycline addition, we observed a substantial elevation in IP 6 (Fig. 5D). Taken together, these data suggest that IPK2 and IPK1 activities are key determining steps in the production of IP 5 and IP 6 in Rat-1 cells.
Down-regulation of Rat IPK2 Using Vector-based siRNA-The data presented so far are from overexpression of genes, that is to say the effects of the gain of function. Another way of analyzing the function of gene products is by eliminating their expression to see the effects of the loss of function. To do this, we applied RNAi (32,36,37). To avoid experiment-to-experiment variability due to different transfection efficiencies with the in vitro synthesized siRNAs, we used a vector-based siRNA expression system (32) to knock down the rIPK2 expression. Fig. 6A shows the Northern blotting data using the entire rIPK2 open reading frame as a probe. Cells expressing siRNA against rIPK2 had a 90% decrease of rIPK2 mRNA compared with cells expressing the vector control (Fig. 6A). Control actin amounts were not changed. To determine whether the knockdown of rIPK2 effected inositol polyphosphate production, we labeled these cells using [ 3 H]inositol, and the soluble inositol polyphosphates were isolated and analyzed by HPLC. The cells expressing siRNA against rIPK2 had decreased I(1,3,4,5,6)P 5 levels to about 20 -25% of the control cells (Fig. 6, B and C). The rIPK2 siRNA did not effect IP 6 levels.
Overexpression of IPK2 or IPK1 Activities Is Not Sufficient to Induce Anchorage-independent Growth of Rat-1 Cells-Several studies have indicated that transformation of Rat-1 cells with the oncogene v-Src results in activation of inositol trisphosphate kinase activity and elevation of IP 4 levels (38 -40). We therefore tested whether or not gain of function of either IPK2 or IPK1 activities in Rat-1 cells enabled anchorage-independent growth (Fig. 7). Overexpression of either scIpk2 or rIPK2 increased IP 4 and IP 5 levels in cells; however, soft agar analysis demonstrated that these modified cells were not able to grow (Fig. 7, B and C). In contrast v-Src-transformed Rat-1 cells were able to grow in soft agar (Fig. 7F). We next examined Rat-1 cells stably overexpressing either scIpk1 or rIPK1 in soft agar assays (Fig. 7, D and E). Despite the elevated levels of IP 6 neither cell line was able to divide in an anchorage-independent manner. , pRevTRE/GFP-rIPK2 (E), pRevTRE/GFP-rIPK2 D127A (F), pRevTRE/GFP-ScIPK1 (G), or pRevTRE/ GFP-rIPK1 (H) were seeded into a 60-mm culture dish containing 3 ml of culture media. One day after growth, cells were labeled in Medium-199 supplemented with 10% dialyzed FBS and containing 20 Ci/ml [ 3 H]inositol for 2 days. 2 g/ml doxycycline was added to the media to induce protein expression for the same period. The soluble inositol polyphosphates were extracted and then analyzed by HPLC using the Partisphere strong anion exchange column as described under "Experimental Procedures." The elution positions of inositol polyphosphates were compared with the elution of known species and are indicated by the arrows. Ins, inositol.

DISCUSSION
There are several reports that have described functions for the highly phosphorylated inositol polyphosphates such as IP 5 and IP 6 in eukaryotic cells (1)(2)(3). However, the molecular route of their synthesis has remained unclear in higher eukaryotic cells. To elucidate the molecular basis for synthesis of these important messengers we established cell lines in which inositol polyphosphate levels were altered through overexpression or RNAi of inositol polyphosphate kinases. We found that rat IPK2 and rat IPK1 play a critical role in the synthesis of IP 5 and IP 6 . Our data suggest that rIPK2 and rIPK1 activities are rate-determining steps in rat IP 5 and IP 6 production and appear highly regulated. Importantly we demonstrated, both in vitro and in vivo, that rIPK2 and rIPK1 activities are sufficient to convert I(1,4,5)P 3 to IP 6 and thus are the minimal two kinases required for IP 6 synthesis in rat cells. This observation is consistent with what we have reported in budding yeast, plants, and flies.
Our data indicate that absolute levels of IP 6 in Rat-1 cells are tightly controlled such that it appears that feedback mechanisms sense and control its levels. The simultaneous overexpression of rIPK2 and rIPK1 where the level of IP 6 appears to reach an upper threshold at which point it may be degraded support this hypothesis. For instance, the IP 6 levels in the IPK2-and IPK1-overexpressing cells did not proportionally elevate the levels of IP 6 compared with IPK1-overexpressing cells. In addition, decreasing cellular I(1,3,4,5,6)P 5 levels using siRNA against rIPK2 significantly altered the ratio of IP 5 to IP 6 . This indicates that rIPK1 activity has been up-regulated to compensate for IP 6 loss.
The substrate specificity of mammalian orthologs of IPK2 is variable according to recent reports (10 -12). Our data showed that rIPK2 phosphorylates I(1,4,5)P 3 to I(1,3,4,5,6)P 5 through I(1,3,4,5)P 4 in vitro. Chang et al. (12) have reported that the human IPK2 ortholog acts primarily as a I(1,3,4,6)P 4 5-kinase using kinetic studies and yeast complementation assays, which showed that the human ortholog did not complement inositol polyphosphate production in ipk2⌬ yeast. We found that rIPK2 was able to complement IP 5 and IP 6 production in ipk2⌬ yeast demonstrating its ability to convert I(1,4,5)P 3 to I(1,3,4,5,6)P 5 . The different specificities may arise from slight sequence variation as rIPK2 is 83.6% identical to human IPK2 (12). Reports of two other groups have indicated that human IPK2 is capable of converting I(1,4,5)P 3 , although these studies only examined activity in vitro and did not provide proof in cells (10,11). It is also noteworthy that the IP 4 intermediate of I(1,4,5)P 3 to I(1,3,4,5,6)P 5 phosphorylation by rIPK2 was different from that of budding yeast, plant, and fly Ipk2, all of which prefer to first phosphorylate the D-6 position of I(1,4,5)P 3 and then to phosphorylate the D-3 position to produce I(1,3,4,5,6)P 5 (6,7,29,41). Moreover the rIPK2 D127A mutant, which corresponds to the Scipk2 kin Ϫ (D131A) mutant (6), still showed weak but significant activity toward I(1,4,5)P 3 both in vitro and in vivo. Thus it appears that subtle changes in amino acids in the active site allow for significant alterations in substrate recognition between rat and yeast Ipk2. This feature will likely be highly useful for future studies aimed at designing substrateselective Ipk2 mutants.
Are there multiple pathways in place that enable IP 6 synthesis in mammalian cells? Given the importance of these IP regulators, it is likely that alternate pathways may exist. Of interest, among the pathways proposed (see Fig. 8) the last two steps require IPK2 and IPK1 activities. In mammalian cells, one proposed alternate route involves two additional kinases and one phosphatase: 1) an IP 3 3-kinase (which based on sequence similarity is a more recently evolved member of the IPK2 superfamily); 2) an I(1,3,4)P 3 5/6-kinase, which generates either I(1,3,4,5)P 4 or I(1,3,4,6)P 4 ; and 3) an I(1,3,4,5)P 4 5-phosphatase. Whether or not this route plays a role in IP 5 and IP 6 synthesis in Rat-1 cells is not known. Future studies of the rat I(1,3,4)P 3 5/6-kinase will be important for determining this in Rat-1 cells. Although we clearly demonstrated in this study that rIPK2 is capable of bypassing these alternate three steps and directly phosphorylates I(1,4,5)P 3 to I(1,3,4,5,6)P 5 via I(1,3,4,5)P 4 in yeast and Rat-1 cells. However, since our RNAi studies indicated that loss of rIPK2 diminishes but does not eliminate IP production, it is possible that the alternate pathway is contributing in part to IP 6 production. It is also possible since the RNAi only partially knocked down rIPK2 that this accounts for the residual IP 6 synthesis. It is also tantalizing to In S. cerevisiae, Arabidopsis, and D. melanogaster, I(1,4,5)P 3 is phosphorylated predominately first on the D-6 position to generate I(1,4,5,6)P 4 and then the D-3 position to generate I(1,3,4,5,6)P 5 by Ipk2 (designated ScIpk2 for the original characterization in S. cerevisiae). Our studies of rat IPK2 indicate that it prefers to convert I(1,4,5)P 3 to I(1,3,4,5,6)P 5 via an I(1,3,4,5)P 4 intermediate. IP 3 3-kinase, a recently evolved relative IPK2, also generates I(1,3,4,5)P 4 as its sole product. However, given the data presented in this report, previous published work in D. melanogaster (41), and mouse knock-out data of others (46 -48), it does not appear that IP 3 3-kinase (IP 3 3-K) contributes to the direct synthesis of IP 5 and IP 6 in flies, rat, or mouse cells. Similar to yeast, plant, and flies, the rat IPK1 acts as a 2-kinase to convert I(1,3,4,5,6)P 5 to IP 6 . Finally our work does not exclude the involvement of I(1,3,4,5)P 4 5-phosphatase (5-ptase) and I(1,3,4)P 3 5/6-kinase (IP 3 5/6-K) activities in IP 6 production in mammals; however, it does demonstrate that rIPK2 and rIPK1 activities are sufficient to convert I(1,4,5)P 3 to IP 6 in cells. speculate that the other reported roles of the I(1,3,4)P 4 5/6kinase, as a regulator of protein phosphorylation (42,43) and I(3,4,5,6)P 5 metabolism (44), may function to control IP metabolism independently of I(1,3,4)P 3 phosphorylation. Given that rIPK2 functions as an I(1,4,5)P 3 3-kinase in vivo, it remains an important question to determine the functional redundancy or difference between IPK2 and IP 3 3-kinases, which are only capable of phosphorylating the D-3 position of I(1,4,5)P 3 (2). Do these two enzymes compete with each other for substrate, or are they compartmentalized to access different pools of substrate in the cells? Three molecular studies indicate that IP 3 3-kinases are not involved in IP 5 and IP 6 synthesis. 1) Balla et al. (45) suggested that overexpression of IP 3 3-kinase did not increase higher inositol polyphosphates such as IP 5 and IP 6 in NIH 3T3 cells; 2) IP 3 3-kinase mouse knock-out cells reveal that IP 5 and IP 6 production is unaltered (46 -48); and 3) in Drosophila melanogaster loss of IP 3 3-kinase isoforms does not decrease IP 6 levels, but loss of dmIpk2 nearly ablates IP 5 and IP 6 (41).
Several reports have suggested the involvement of highly phosphorylated inositol polyphosphate production in cell proliferation and cell transformation (38, 49 -51). We could not detect a significant difference of proliferation rate and colony formation ability among control and our inositol polyphosphate kinase overexpression and IPK2 knock-down Rat-1 cells (not shown). Thus it appears that 2.5-fold elevation or 80% reduction of IP 5 does not sufficiently alter growth. It is likely that the remaining 20% activity in the RNAi knock-down cells is sufficient for survival; thus the examination of IPK2 or IPK1 complete knock-outs will be important future studies. Additionally investigating the relationship between the production of highly phosphorylated inositol polyphosphates and IPK2 and/or IPK1 activity during the cell cycle, cell differentiation, proliferation, and transformation may be equally important.
In conclusion, in this study we established an in vivo model system to modulate the inositol polyphosphate levels. Importantly we found that, in Rat-1 cells, rIPK2 and rIPK1 are the minimal kinase activities required to synthesize IP 6 from I(1,4,5)P 3 via I(1,3,4,5)P 4 and I(1,3,4,5,6)P 5 intermediates. This now confirms that in S. cerevisiae (5,6), Arabidopsis (29), D. melanogaster (41), and now Rattus norvegicus these two kinases are functionally important. The ability to modulate the intracellular inositol polyphosphate levels will provide powerful tools to study the roles of I(1,3,4,5,6)P 5 and IP 6 in eukaryotic cell signaling.