Short Communication Inositol Tetrakis- And Pentakisphosphates In Gh₄ Cells

The attention of many laboratories has been focused upon the role of inositol 1,4,5-trisphosphate (Insl,4,5P3) as a specific releaser of intracellular calcium (Berridge, 1984; Berridge & Irvine, 1984). In addition to Insl,4,5P3, stimulated cells also produce Insl,3,4P₃ (Irvine, Letcher, Lander & Downes, 1984; Irvine, Anggard, Letcher & Downes, 1985; Burgess, McKinney, Irvine & Putney, 1985) but neither the function nor the source of this isomer is known. Batty, Nahorski & Irvine (1985) have proposed that the precursor of In8 1,3,4P₃ is inositol 1,3,4,5-tetrakisphosphate (Ins P₄). Here we confirm the existence of Ins P₄ in a pituitary cell line (GH₄ cells) and describe, in addition, the existence of both inositol pentakisphosphate (Ins P₅) and inositol hexakisphosphate (Ins _P6, phytic acid). We have also measured changes in these various inositol phosphates following stimulation with thyrotropin-releasing hormone (TRH).

To test the effect of stimulation, GH₄C₁ cells were grown in 100-mm dishes with 8 ml of Eagles MEM supplemented with 11% foetal calf serum and 72 μCi of myo 2-³H-inositol (Amersham) per dish for about 86 h. Cells were harvested in calcium-free assay medium plus 0·02% EDTA at 37°C for 5 min, washed twice in assay medium (92 ml of 1·28 mol 1^–1 NaCl, 5 ml of 2 mol 1^–1 KC1, 5 ml of 0·2 mol 1^–1 CaC12, l·8g of D+ glucose, 4·8g of HEPES per litre, pH7-2 with NaOH) and resuspended in assay medium. Each 100-mm dish provides cells for 10 assays. The cells were stored in a plastic tube in crushed ice and gently agitated before sampling. 100μl of mixed cells were warmed to 37°C and treated with 200μl of TRH (final concentration 1·0μmoll^–1). The reaction was stopped with 300μl of 10% trichloracetic acid (TCA) and the cells allowed to extract for 5–10 min in ice. To reduce the risk of phosphate migration on InsP₄, the aqueous extracts were immediately washed three times with 10 volumes of ether to remove TCA and stored at — 18°C. Before chromatography, they were filtered through a 0·45 μl m filter (ACRO LC13, Gelman Sciences). Successive applications of hormone and TCA at short intervals were made using an apparatus described elsewhere (Berridge, Buchan & Heslop, 1984).

Separation of the inositol phosphates was achieved by HPLC (Partisil 10 SAX columns -Technicol Ltd, Higher Hillgate, Stockport, Cheshire) using a modification of the gradient system described by Irvine et al. (1985). The flow rate was 2·5 ml min^–1 and fractions were collected at 30-s intervals using two linear ammonium formate gradients as described in Fig. 1. Scintillation counting of the column eluates was by adding 9 ml of ACS counting fluid (Amersham International, Bucks) for all vials up to fraction no. 59, i.e. the Ins 1,4,5Ps and Ins 1,3,4P3 peaks. For the fractions containing the more polar compounds (fractions 60–85) methanol was added to the scintillation fluid to prevent phase separation (Irvine et al. 1985). 10 ml of a mixture of ACS and 50% methanol (5:1) was added to 0·5 ml of each fraction.

The putative InsP₄, Ins P₅ and Ins P₆ were desalted (Irvine et al. 1984) before analysis by ionophoresis in 0·1 mol 1^–1 Na oxalate pH 1·(Seiffert & Agranoff, 1965) with internal standards of Ins P₃, InsP₄, InsP₅ and InsP₆ prepared by acid hydrolysis of phytic acid (Desjobert & Petek, 1956). The putative Ins P₅ was also dephosphorylated by 6 mol 1^–1 HC1 (110°C, 18 h) and the resulting radiolabelled sugars were analysed by ionophoresis in 0·1 moll^–1 NaOH (Frahn & Mills, 1959).

The separation of inositol phosphates extracted from GH₄ cells is shown in Fig. 1. The two isomers of Ins P₃ run close together but the other three peaks are widely separated from each other. These three late peaks were analysed by ionophoresis and were found to have identical mobilities to (in order of elution) InsP₄, Ins P₃ and Ins P₃. The first of these was not analysed further, but it is likely to be the same compound, Ins 1,3,4,5P₄, described by Batty et al. (1985). The latter two are novel compounds in mammalian tissues; the putative InsP₅ was found to give inositol as the only detectable radioactive product. From this, and from their ionophoretic mobilities we identify them as Ins P₅ and Ins P₆, though this is as yet a preliminary identification which requires further confirmation and isomer determination.

Extracts prepared from ³H-inositol-labelled mouse Swiss 3T3 cells, locust (Schistocerca gregaria) eye and blowfly (Calliphora vomitoria) salivary gland (Table 1) were also found to contain peaks corresponding to Ins P4, Ins P₅ and Ins P₆. Taken together with the report of Ins P4 in rat cerebral cortex (Batty et al. 1985) it seems that all three compounds are widely distributed in animal cells.

In order to determine whether these higher inositol phosphates play a role in stimulus-response coupling we studied how their levels varied after addition of TRH to GH₄ cells (Fig. 2) or 5-hydroxytryptamine (5-HT) to blowfly salivary gland (Table 1). The most pronounced change in GH₄ cells was an increase in the level of Ins 1,4,5P3 with a delayed increase in Ins 1,3,4P3 which shows that the cells were properly stimulated (Fig. 2). There were also small but rapid transient fluctuations in Ins P₄ and Ins P₅ which appeared to occur in unison. An initial increase at 0·5 s was followed by a very significant fall. At subsequent times the levels oscillated around the resting level. The InsP₄ level was elevated above the control value at 5 min but there was no change in either Ins P₅ or Ins P₆. The latter did not change at any time period. However, the amount of Ins P₆ seemed to vary between batches of GH₄ cells and work is in progress to resolve this anomaly.

The results obtained following a 10-s stimulation of blowfly salivary glands are shown in Table 1. Not only was the now conventional increase in Ins 1,4,5P3 found but in this tissue there is a substantial increase in Ins 1,3,4P3 even at this early time. A five-fold increase in the label in Ins P4 underlines the probable importance of this intermediate while the marginal reductions in Ins P₅ and Ins P₆ remind us that a role for these compounds in the overall metabolic response to stimulation cannot be excluded from consideration in future work.

There are marked differences between the changes in InsP₄ shown in Fig. 2 and the much larger increases in radioactivity in the same compound seen in blowfly salivary glands (Table 1) and reported by Batty et al. (1985) in rat cerebral cortex slices. A common feature of the last two experiments is the shorter labelling times that were used (4 h at room temperature for the salivary glands and 1 h at 37°C for the brain slices). It might be argued that GH₄ cells ought to approach uniformity of labelling in 86 h at 37°C and that during shorter periods of inositol incorporation there remained a disparity between the specific activities of inositol compounds in different pools. The argument to fit the data so far available would be that the inositol lipid pool had the higher specific activity before stimulation and that the sharp rise in radioactivity in InsP₄ following stimulation was due less to a change in the concentration of InsP₄ than to a rise in its specific radioactivity. An equally convincing explanation, however, would be to point to the evidence that in GH₄ cells Ins P₅ seems to be in rapid equilibrium with Ins P₄ and that the two intermediates taken together form a larger chemical reservoir. Moreover, the rapid changes in radioactivity of both compounds suggest that they, in turn, must be converted to further metabolites. No interpretation of the data will be convincing unless the source of Ins P₄, InsP₅ and InsP₆ can be clearly established.

One obvious possibility is that each inositol phosphate is derived from a corresponding inositol lipid (Fig. 3). Batty et al. (1985) have already proposed that Ins?4 originates from phosphatidylinositol 3,4,5-trisphosphate (PIP3). There already is some evidence for such a lipid in animal cells (Santiago-Calvo, Mulé & Hokin, 1963) though the data of Seiffert & Agranoff (1965), and Batty et al. (1985) suggest that if it exists, it does so at very low levels. Perhaps membranes also have small amounts of PIP4 and PIP5 which when hydrolysed will yield Ins P5 and Ins Pô (Fig. 3). An alternative is that these higher inositol phosphates are derived through the action of inositol phosphate kinases which are known to exist in both plants and avian erythrocytes (Chakrabarti & Majumber, 1978). The latter are known to contain large amounts of Ins 1,3,4,5,6P₅ which performs a role similar to that of 2,3-phosphoglyceric acid in stabilizing haemoglobin. These avian erythrocytes contain an inositol phosphate kinase capable of phosphorylating Ins P₁ to Ins P₆ through the sequence of reactions shown in Fig. 3. It seems that animal cells also possess the kinase reactions necessary to make all the inositol phosphates. Whether these kinases act on inositol which is free or that which is attached to lipid remains to be seen.

Finally, it is interesting to consider whether these novel inositol phosphates may function as second messengers. Batty et al. (1985) have suggested that InsP₄ may have a second messenger role in calcium mobilization similar to that previously proposed for In8 1,4,5P₃ (Berridge & Irvine, 1984). There was a delay of approximately Is preceding the large increase in Insl,4,5P3 in GH4 cells (Fig. 2) which is rather long in the light of previous Quin2 measurements indicating that intracellular calcium levels begin to rise in considerably less than 1 s (Albert & Tashjian, 1984). It is conceivable, therefore, that if InsP₄ functions ‘as an adjunct to Ins 1,4,5P3 in calcium mobilization’ as Batty et al. (1985) have proposed, then the former might be responsible for very early calcium signalling. There were indications of rapid cyclical changes in InsP4 and Ins P₅ with an initial rise which preceded the production of Insl,4,5P₃ (Fig. 2). Before speculating further on possible second messenger functions, it will be essential clearly to establish the metabolic pathways responsible for generating these inositol phosphates. It has been assumed that these higher inositol phosphates were restricted to plant cells and avian erythrocytes; it comes as a considerable surprise, therefore, to find that they are widely distributed in animal cells. Now that attention has been focused upon them we can expect a rapid advance in our knowledge of their metabolic role. The possibility that different tissues may employ a common underlying mechanism in different ways gives added interest to an intriguing development.

We gratefully acknowledge the generous loan of tissue culture facilities by the Cancer Research Campaign Mammalian Cell DNA Repair Research Group.