INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

INOSITOL PHOSPHOLIPIDS have been recognized as important regulators of a variety of cellular functions. The best characterized of their regulatory roles is the receptor-mediated production of the second messengers, D-myo-inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃] and diacylglycerol, by phospholipase C (PLC)-mediated hydrolysis of membrane phosphatidylinositol-4,5-bisphosphate [PtdIns(4,5)P₂] (4). More recently, it has become evident that PtdIns(4,5)P₂ and its 3-phosphorylated product, PtdIns(3,4,5)P₃, may have additional regulatory roles by controlling the assembly and activities of several protein-signaling complexes at specific membrane compartments (35). This latter aspect of inositol lipid-based signaling makes use of the various forms of inositol lipid kinase and phosphatase enzymes that have been described in recent years (12).

Phosphatidylinositol (PI) 4-kinases are the enzymes that catalyze the 4-phosphorylation of PI, the first step in a reaction sequence that leads to the formation of most of the polyphosphoinositides [recent evidence suggests that some 5-phosphorylation of PI may precede 4-phosphorylation (30)]. The majority of the cellular PI 4-kinase activity is represented by the tightly membrane-bound type II 4-kinase, an activity that has been purified from various tissues as a 50-56-kDa protein (see Ref. 6) but still awaits molecular cloning and characterization. In contrast, two distinct forms of the less abundant type III phosphatidylinositol 4-kinases (PI4KIII) have been purified from bovine adrenal (2) and brain (14) and cloned from various species, including humans (23). These two enzymes, the 110-kDa  form [called 92-kDa PI 4-kinase in the rat based on its calculated molecular size (25)] and the 210-kDa  form [called 230-kDa PI 4-kinase in the rat (24)] are homologues of two yeast PI 4-kinases, products of the PIK1 and STT4 genes, respectively (10, 39). Also, these proteins contain the characteristic signature of the ATP-binding catalytic domain of PI 3-kinases and PI kinase homologues (17) and are also inhibited by the microbial product, wortmannin, the most potent inhibitor of PI 3-kinases (37).

It is an intriguing question as to why these two PI 4-kinases evolved so early and remained highly conserved during evolution [the homologues of both enzymes have also been cloned from plants (33, 38)]. Genetic evidence of the importance of these proteins in yeast indicates that both proteins are essential for survival in most strains (7, 10) and that they cannot substitute for one another's function. In the present study, we compared the tissue distribution of these two enzymes by in situ hybridization in the rat to obtain further clues about their relative importance in the various tissues. Our results show that although both enzymes are ubiquitously expressed, there are notable differences in their tissue distribution, and that both the smaller , and the larger  forms are most prominently expressed in neurons.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Riboprobes were made from PCR products that were generated with primers in which the T3 (for the sense) and T7 (for antisense) RNA polymerase recognition sequences were added to flank the upstream and downstream primers, respectively. For PI4KIII, sequences from 5245-5738 of the rat PI 4-kinase p230 were used, and for PI4KIII, sequences from 2465-2772 of the rat PI 4-kinase p92 were used [see (1) for nomenclature]. At these regions there was low homology between the nucleotide sequences of the two enzymes (~50%). Riboprobes were made from the gel-purified PCR products as templates using ³⁵S-labeled UTP and the respective RNA polymerase.

For in situ hybridization histochemistry, a protocol displayed on the website (http://intramural.nimh.nih.gov/lcmr/snge) was followed (19). Briefly, 12-µm tissue sections made from frozen tissue blocks obtained from male Sprague-Dawley rats (killed under CO₂ anesthesia for other experimental purposes) were fixed with 4% paraformaldehyde and washed twice with RNAse-free PBS. After this treatment, the slides were subjected to 0.25% acetic anhydride (freshly made in 0.1 M triethanolamine/HCl, pH 8.0) for 10 min followed by sequential washes in increasing concentration of ethanol.

Tissue sections were hybridized in a wet chamber with the radioactive probes for 22 h at 55°C. Slides were rinsed with 4× sodium chloride-sodium phosphate-EDTA (SSPE)/1 mM 1,4-dithiothreitol (DTT) four times for 5 min at room temperature before two 30 min washes of 65°C with 0.1× SSPE/1 mM DTT. Before drying, slides were rinsed twice with 1× SSPE at room temperature. The slides were then exposed to X-ray films and subsequently coated with Kodak NTB-3 nuclear track emulsion for 1 wk exposure.

The same riboprobes were used to hybridize for Northern blot analysis of membranes containing poly(A)⁺-selected mRNA from various rat tissues (Clontech) using the method previously described for the human tissues (2).

For comparison of the relative amounts of the two proteins expressed in the brain, two rat brains were homogenized, and the proteins purified, on heparin and MonoQ columns as described previously (2). [³H]wortmannin binding (2) was used to quantitate the proteins from the active fractions.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Distribution of  and  forms of PI4KIII in the rat brain. To characterize the riboprobes chosen for in situ hybridization, we performed Northern blot analysis with the antisense riboprobes on poly(A)⁺-selected mRNA of various rat tissues. The expression patterns and the size of the two transcripts (7.2 kb and 3.3 kb for PI4KIII and PI4KIII, respectively) were found to be almost identical to those published for these two enzymes in the rat (24, 25). Because both  and  forms show relatively high expression in the rat brain (Fig. 1), first we studied the distribution of the transcripts in brain tissue. When a series of coronal sections of rat forebrain were analyzed by in situ hybridization, prominent labeling was observed with the antisense probe of the type III -enzyme, which was mostly confined to the gray matter (Fig. 2, A-E). Neurons of the cerebral cortex, as well as of the olfactory lobe and the hypothalamus, were strongly positive (Figs. 2 and 3). No signal was detected in any of the brain areas with the sense probes (see Fig. 2 for an example). The strongest signal was detected in limbic areas such as the cell bodies of pyramidal cells in the CA1-CA3 layer of the hippocampus and the granule cells of the dentate gyrus (Fig. 2, B and C). The amygdaloid nucleus and the entorhinal cortex also showed a very strong signal, indicating the abundant presence of mRNA for PI4KIII.

View larger version (100K): [in this window] [in a new window]   Fig. 1.   Northern blot analysis performed with riboprobes used for in situ hybridization studies on a rat poly(A)+-selected RNA panel (Clontech). Signals that were obtained with antisense probe against rat homologue of PI4KIII [top; named 92-kDa phosphatidylinositol (PI) 4-kinase (25)], and antisense probe against rat PI4KIII [bottom; named 230-kDa PI 4-kinase (24)] are shown. Size of RNA marker is indicated, left. Sk, skeletal muscle.

View larger version (99K): [in this window] [in a new window]   Fig. 2.   In situ hybridization analysis performed on selected coronal sections of adult rat brain. Results obtained with antisense (A, C-F, H-J) or sense (B and G) riboprobes for PI4KIII and PI4KIII, respectively, are shown. Dark-field images on sections through frontal cortex and olfactory bulb (5.6 mm anterior to bregma level for A, B and F, G), thalamus (3.5 mm posterior for C and H and 5.0 mm posterior for D and I), and cerebellum (9.9 mm posterior to the bregma level for E and J) are shown. Note intense signal above mitral cells (Mi) and internal granular (IGr) layer of olfactory bulb with both antisense (but not sense) probes. Strong signal was also found above neurons in limbic areas (hippocampus, amygdala), and cerebellum. Cpa, parietal cortex; Hi, hippocampus; Ic, internal capsule; Th, thalamus; Ha, habenula; Am, amygdala; Ch, choroid plexus; Sb, subiculum; Ce, entorhinal cortex; Cp, cerebral peduncle; Ml, molecular layer; Gl, granular layer; Lc, locus ceruleus; Ntd, dorsal tegmental nucleus.

View larger version (188K): [in this window] [in a new window]   Fig. 3.   In situ hybridization analysis performed on a coronal section of adult rat brain via hypothalamus (2.1 mm posterior to bregma level) with antisense PI4KIII (A, C, and E) and PI4KIII (B, D, and F) riboprobes. Dark-field (A-D) and bright-field (E and F) images are shown demonstrating strong positive signals in paraventricular (Pvn) and supraoptic (Son) nuclei with both probes. Note diffuse signal in white matter and staining of ependymal lining of 3rd ventricle (3rdV) with antisense - but not -probe (B, D, and F).

PI4KIII is the predominant enzyme in the brain. To compare the relative amounts of the two enzymes expressed in the brain, rat brains were homogenized and the membrane fraction solubilized with cholate essentially as described by Endemann et al. (9). The wortmannin-sensitive PI 4-kinase activity was enriched by heparin and MonoQ chromatographies. Given the comparable affinities of the two proteins for wortmannin (2), [³H]wortmannin binding was used to assess the relative quantities of the two proteins from the active fractions. As shown in Fig. 4, the larger -enzyme represented the majority of the activity in the rat brain (as well as in the bovine brain, not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Relative abundance of PI 4-kinase type III  and  proteins enriched from rat brain. Membrane fractions obtained from rat brain homogenates were solubilized with cholate (9) and their wortmannin-sensitive PI 4-kinase activity enriched by heparin and MonoQ chromatographies (2). [3H]wortmannin binding was performed on active fractions followed by SDS-PAGE analysis and autoradiography. Most of type III PI 4-kinase activity (>90%) is attributed to larger enzyme. Mr, relative molecular weight.

Both PI4KIII and PI4KIII are expressed in the retina. Next we examined whether the transcripts for these two enzymes were detectable in the retina. As shown in Fig. 5, a strong signal was detected above the ganglionic cells, especially with the antisense -probe, and the inner nuclear layer. A somewhat weaker signal was found in the pigment epithelium and in the outer nuclear layer. The signals were generally weaker with the -probe than with the -probe. Again, only a low evenly distributed background signal was detected with either of the sense riboprobes (Fig. 5).

View larger version (159K): [in this window] [in a new window]   Fig. 5.   In situ hybridization analysis performed on sections of retina of adult rat eyes with antisense (A, B and E, F) and sense (C, D and G, H) riboprobes against rat PI4KIII (A-D) or PI4KIII (E-F). Bright-field (A, C, E, and G) and dark-field (B, D, F, and H) images of same sections are shown. All layers containing cell bodies (G, ganglion cells; I, internal molecular layer; O, outer molecular layer; P, pigment cell layer) showed positive signal with both antisense probes.

Localization of the two PI 4-kinase enzymes in peripheral tissues. Based on Northern blot analysis, the abundance of mRNA for the  form of PI 4-kinase was generally very low but detectable in peripheral tissues. In contrast, PI 4-kinase  was more widely expressed (Fig. 1 and Refs. 24 and 25), and, therefore, its distribution was further examined by in situ hybridization histochemistry in peripheral tissues. As shown in Fig. 6, PI4KIII mRNA was detected in the kidney, especially in the papilla of the medulla, and also in the glomeruli of the cortex. In the spleen there was a diffuse signal both over the red and the white pulp. In the heart, the signal was confined to the atria, the ventricles showing very low if any expression. The testis showed a clear signal over the seminiferous tubules, but little in the interstitium. Of the other tissues tested, a weak signal was found in the liver and the stomach (not shown).

View larger version (120K): [in this window] [in a new window]   Fig. 6.   In situ hybridization of tissue sections obtained from rat kidney (A-D), spleen (E, F), heart (G, H) and testis (I, J) with riboprobes against rat PI4KIII. Bright-field (left) and dark-field (right) of same sections are shown. More prominent signal was obtained above kidney papilla (Pa) and glomeruli (Gl). A more diffuse signal was detected in spleen (F; Wp, white pulp; tv, trabecular vein). In heart, only atrium (Atr) but not ventricles (Ven) showed a positive signal. In testis, seminiferous tubules (S) but not interstitial cells (I) showed expression of enzyme.

Localization of the two PI 4-kinase enzymes in rat embryo. The expression pattern of the two PI 4-kinase enzymes was also examined in a midsagittal section of a 17-day-old rat embryo (Fig. 7). The  form of the enzyme showed high expression levels in the developing brain and eye, as well as in the trigeminal ganglia. Among the peripheral tissues, a relatively strong signal was detected in the salivary glands, the lungs, and the liver. The  form of the enzyme was again more evenly expressed, but the brain and trigeminal ganglion showed the highest expression levels.

View larger version (158K): [in this window] [in a new window]   Fig. 7.   In situ hybridization of sagittal sections obtained from 17-day-old rat embryo with riboprobes against rat PI4KIII (A) or PI4KIII (B). Dark-field images show prominent neuronal localization of PI4KIII and a more even tissue distribution of PI4KIII. Both probes labeled strongly the salivary gland (Sg) and its duct, and the trigeminal ganglion (Tg). Lv, liver; Lu, lung; E, eye.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The present study was undertaken to investigate the tissue distribution of the two cloned type III PI 4-kinase enzymes, the  and  forms, to complement similar studies on other important protein members of the inositide-Ca²⁺ signal transduction cascade (13, 18, 31). Both of these enzymes were found to be widely expressed in the rat, but the  form was predominantly localized to the brain in both the fetus and adult animals. Although more uniformly expressed, the  form of the enzyme also showed highest expression in the brain. The brain still contained mostly the  form and a relatively small amount of the  form of the enzyme when the protein amounts were compared. The brain has long been known to contain the highest concentrations of polyphosphoinositides (11), and both type II and type III PI 4-kinase activities have been isolated from brain tissue (9, 14). In the present study, both type III enzyme mRNAs showed the highest levels of expression in the hippocampus and dentate gyrus, areas that are generally rich in proteins involved in inositide/Ca²⁺ signaling. In most part, the gray matter showed a clear signal with both riboprobes, and the major difference between the distributions of the mRNAs encoding the two proteins was the presence of only the  form of the enzyme in the white matter (in glial cells) and the choroid plexus. These results suggest that the functions of the two enzymes cannot be distinguished based on tissue attributes, and that very likely they subserve nonredundant functions probably within the same cell. This finding is consistent with the separation of these two enzymes early in evolution (10, 33, 39).

The soluble type III PI 4-kinases have been described as targets of the PI 3-kinase inhibitor, wortmannin, and it was found that their inhibition leads to the rapid loss of receptor-regulated PtdIns(4)P and PtdIns(4,5)P₂ pools in several types of cells labeled with either myo-[³H]inositol or [³²P]phosphate (27). Based on such evidence it was postulated that one (or both) of these enzymes is responsible for the maintenance of agonist-sensitive phosphoinositide pools (27). Receptor-regulated generation of inositol lipid-based second messengers is one of the most fundamental signaling mechanisms that regulates a great variety of cellular responses. This process has been found to transmit signals from many classes of neurotransmitter receptors present in the brain, such as the muscarinic, -adrenergic, serotoninergic, and metabotropic glutamate receptors. Lithium ions are inhibitors at specific dephosphorylating steps of inositol-phosphate metabolism and hence prevent efficient recycling of myo-inositol in the brain, which relies on its own inositol pool. It has been proposed that the therapeutic effect of Li⁺ in the treatment of manic-depressive disease is related to the ability of this ion to affect PI turnover (5).

Consistent with the importance of inositol lipid-based postreceptor signaling in the brain, all major isoforms of PLC (, , and ) as well as the Ins(1,4,5)P₃ receptor have been found in the brain (32), and their mRNA distributions have been determined (18, 26, 31, 36). However, recent advances in research on neurotransmitter release indicate that inositides may also participate in presynaptic events in addition to their above-discussed signaling role (8). Among the proteins identified as critically important in neurotransmitter release, both the GTPase protein, dynamin (34), and the phosphoinositide phosphatase, synaptojanin (22), have been linked to inositides. Dynamin has a pleckstrin homology domain, which is believed to confer regulation by PtdIns(4,5)P₂ to the protein, whereas synaptojanin is a 5-phosphatase enzyme that, like dynamin, associates with amphiphysin and undergoes dephosphorylation, presumably regulating the level of PtdIns(4,5)P₂ (reviewed in Ref. 8). Intriguingly, a recent report identified Pik1, the yeast homologue of PI4K, as a binding partner for yeast Frq1, a yeast homologue of neuronal frequenin (16). Frequenin, a member of the group of small Ca²⁺-binding regulatory proteins, has been found to modulate synaptic efficacy in neurons in Drosophila (29). These recent findings confirm that inositol lipids possibly play a pivotal role in neurotransmitter release and regulated secretion (21). Expression of both forms of type III PI 4-kinases in the brain is consistent with the importance of inositol lipid production and metabolism in the regulation of these processes.

Given the prominent neuronal localization of both forms of type III PI 4-kinase, the presence of both mRNAs above the neuronal layers of the retina is not surprising. However, their presence above the neurons of the outer nuclear layer, where the cell bodies of the photoreceptors are located, is of particular interest. Although the role of phosphoinositide-based messengers in invertebrate photo-signal transduction is firmly established, it is not clear if this system plays any role in the vertebrate eye where the cGMP system is the primary signaling mechanism (3). Six PLC isoforms have been isolated from bovine retina yielding to the cloning of a then novel form, PLC4, a putative functional homologue of the Drosophila NorpA (20). The presence of PLC4-like immunoreactivity in the rod outer segment has also been demonstrated (28). These findings suggest that inositide-based signaling may still be integrated into the mammalian photosensory system, perhaps at the level of synapses.

Consistent with the ubiquitous role of inositide-based signaling, we demonstrated the presence of the type III PI 4-kinase, especially the smaller  form, in peripheral tissues, although at a significantly lower level than in the central nervous system. It is noteworthy that a stronger signal was detected with the -probe above the kidney papilla corresponding to the cells of the collecting ducts, and also over the duct of the salivary glands in the fetus. Moreover, the choroid plexus and the ependymal lining of the ventricles also showed a selective signal with the -probe. This raises the possibility that this enzyme may have an important role to play in intracellular processes that participate in fluid transport. Although there are reports on the role of PI 3-kinases (based on inhibitor sensitivity) in transcytosis through epithelial cell layers (15), no such data are available on PI 4-kinases.

Comparison of the results of our previous Northern blot analysis using human mRNA (2; also see Ref. 23) and those of the current in situ hybridization shows some discrepancies. In this regard it is important to emphasize that the Northern analysis performed in the present study with the same riboprobes used for the in situ studies and a commercial mRNA blot (Clontech) showed very similar results to those described earlier in the rat using mRNAs prepared by those investigators (24, 25). Strong signal was previously detected in the heart and skeletal muscle with the -probe on Northern blot analysis of human tissues (2, 23). However, this was not present in the rat, and both tissues showed very low expression of the two enzymes both by Northern analysis and in situ hybridization. The reason for this discrepancy between human and rat tissues is not clear at present.

In summary, the present results demonstrate that type III PI 4-kinase  and  are ubiquitously expressed in various tissues, but show predominant brain localization. Although these data are consistent with the important signaling role of inositol lipids in neurotransmission, they also suggest divergent, nonredundant functions of the two enzymes. Future studies are aimed at defining these divergent roles to better understand the complex regulatory role of inositol lipids in controlling multiple cellular functions.