Advan. Enzyme Regul., Vol. 32, pp. 91-103, 1992
0065-2571/92/$15.00 © 1992Pergamon Press pie.
Printed in Great Britain. All rights reserved
CHANGES IN INOSITOL LIPID METABOLISM AND PROTEIN KINASE C TRANSLOCATION IN NUCLEI OFMITOGEN STIMULATED SWISS 3T3 CELLS LUCIO COCCO*, ALBERTO M. MARTELLI*, R. STEWART GILMOUR§, ROSA ALBA RANAt, OT'I'AVIO BARNABEI:~ and FRANCESCO A. MANZOLI* *Institute of Human Anatomy, University of Bologna, Italy tlnstitute of Human Anatomy, University of Chieti, Italy :[:Department of Biology, University of Bologna, Italy §Department of Molecular and Cellular Physiology, Institute of Animal Physiology and Genetics Research, AFRC, Babraham, Cambridge, UK INTRODUCTION
The regulation of genetic material must be intimately coordinated with cellular as well as extracellular events. Clearly the cell is capable of transducing information from the periphery into the nucleus. An understanding of how this information is transmitted requires a thorough knowledge of both the structural organization of the nucleus and the metabolic pathway of the signalling system. Inositol lipid metabolism has become closely identified with the production of multiple informational molecules (1) and the awareness that a phospholipid can act as signal reservoir has become increasingly accepted (2). Indeed the identification of DAG as a physiological regulator of PKC has provided the first good evidence that lipids themselves are actively involved in signal transduction (3). The interest has been augmented by the evidence that IP3, generated by the PLC specific for PIP2, is able to mobilize calcium ions from intracellular stores; thus, a key role for lipid second messengers, inositol phosphates, is to modulate the fluxes inside the cell of the ubiquitous second messenger, Ca 2+ (4). During mitogenic stimulation a number of events takes place. First of all, there is the specific binding of a growth factor to its own receptor on the plasma membrane. Information dealing with the following early Abbreviations: IP3, inositol-l,4,5-phosphate; DAG, diacylglycerol; ATP, adenosine triphosphate; PI, phosphatidylinositol; cAMP, cyclic-adenosine monophosphate; IGF-I, insulin-like growth factor-I; PIP, phosphatidylinositol-4-phosphate; PIP2, phosphatidylinositol-4,5-phosphate; PA, phosphatidic acid; BrdU, 5-bromodeoxyuridine; PKC, protein kinase C; PS, phosphatidylserine; EGTA, (ethylenedinitrilo) tetraacetic acid; SPH, sphingosine; PLC, phospholipase C. 91
92
L. COCCO,et al.
events indicates that some polypeptide hormones, acting as growth factors, induce a transient appearance of cytoplasmic second messengers such as IP 3 and DAG after binding to their own receptor, whereas other factors, acting through a tyrosine kinase receptor, directly activate protein phosphorylation and finally other hormones use the cAMP as second messenger (5). Therefore, the signalling pathways elicited by cell surface receptors activated by growth factors are more complicated as compared to the signals that activate nuclear receptors. These signals indeed are transmitted by small amphiphilic molecules such as steroids that can freely diffuse through lipid membranes and reach the nucleus (6). Nevertheless, the mechanism by which the mitogenic stimulus generated by polypeptide growth factors is eventually transmitted to the nucleus and whether or not a signalling system exists at the nucleus itself remains almost unknown. The evidence for the occurrence at nuclear level of inositol lipid metabolism (7) and the relationship between this metabolism and cell growth and differentiation (8) has suggested that inside the nucleus a signalling system could exist using polyphosphoinositides as messengers. NUCLEAR INOSITOL LIPIDS AND MITOGENIC STIMULATION WITH IGF-I AND BOMBESIN IN SWISS 3T3 CELLS Phospholipids have been identified in the cell nucleus and their pattern differs from that of the whole cell (9). The evidence that they are components of the chromatin adds to their obvious presence in the outer nuclear membrane (10) and it has been reported that changes occur in the nuclear phospholipids during neoplastic transformation (11). Their possible action has been analyzed in vitro and it has been shown that they can modulate replication and transcription and mediate the attachment of newly synthesized DNA to the nuclear matrix (12, 13). Moreover, it has been shown that isolated nuclei are able to synthesize PA, PIP and PIP 2 and that after induction of differentiation of Friend cells the pattern of inositoi lipid phosphorylation changes at nuclear level in a manner consistent with a decrease in PIP phosphorylation and an increase in PIP 2 synthesis (7). Certainly these pioneering observations pointed to the likelihood that nature has put these unique lipids to some important use inside the nucleus in addition to their functions outside it. Interestingly, it has been shown that these phosphorylative events take place also independently from the nuclear envelope, since the phosphorylation of inositol lipids occurs in nuclear matrix and in purified nuclear lamina (14). In addition, isolated nuclei have been shown to incorporate and phosphorylate exogenous PI delivered by a specific PI-transfer protein (15).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
93
Recently we have shown that isolated nuclei of Swiss 3T3 fibroblasts incorporate ATP into inositol lipids and that changes occur in their metabolism when quiescent cells are stimulated to grow with a mitogenic combination of IGF-I and bombesin. Actually this effect on inositol lipid phosphorylation is mainly due to IGF-I which induces a decrease of incorporation of [T32p] ATP into PIP and PIP 2 whereas bombesin on its own has no effect. The changes are transient in that, although they are still detectable after 15 min, after 1 hr they are no longer significant (16). During routine passages of cells we selected a clone still maintaining contact inhibition but unresponsive to the above mitogens. This provided an opportunity to investigate a possible correlation between the onset of DNA synthesis and the early changes in nuclear polyphosphoinositides previously described (17). The capability to respond to growth factor stimulation has been evaluated by analyzing DNA synthesis by means of BrdU incorporation (18). Only responsive cells after mitogenic stimulation incorporate BrdU as demonstrated by immunofluorescence microscopy. The number of responsive cells which replicate their DNA is directly related to the stimulus given (Fig. 1). IGF-I in combination with bombesin is the more efficient condition, while both IGF-I and bombesin on their own give rise to a lower number of replicating cells. The unresponsive cells do not incorporate BrdU either in the presence of IGF-I and bombesin on their own or in the presence of the mitogenic combination.
80I
Responsive ceLLs • Unresponsive ceLLs
70
60
"0
50
o 40 30 2O l0 m
I
2
3
4
FIG. 1. Effect of IGF-| and/or bombesin on stimulation of DNA synthesis in Swiss 3'13 cells. DNA synthesis was assessed by BrdU incorporation after growth stimulation of quiescent cells with the above growth factors for 20 hr. Cells were labelled for 60 rain with 10 p~MBrdU and 1 ~M fluorodeoxyuridine and then processed for immunofluorescence microscopy (18). The histogram represents the percentage of labelled nuclei. (1) control unstimulated cells; (2) IGF-I 20 ng/ml; (3) bombesin 1 nM; (4) IGF-I plus bombesin.
94
L. COCCO, et al.
Three points are essential for establishing a correlation between nuclear inositol lipids and the onset of DNA synthesis: (i) the purity of nuclear preparation; (ii) the expression of type I IGF receptor in unresponsive cells; (iii) the lack of changes in polyphosphoinositide metabolism in unresponsive cells. Figure 2 shows that, as assessed by electron microscopy analysis on a large number of thin sections, nuclei purified from Swiss 3T3 cells in the presence of non-ionic detergent (16) are highly pure and lack the outer membrane and extranuclear debris. The cytoplasmic enzymatic markers are indeed absent since in nuclear preparation we were dealing with the activity of glucose-6-phosphatase is less than 0.1% of the activity present in whole cell homogenates. Swiss 3T3 fibroblasts which do not replicate under mitogenic stimulation as well as responsive cells express the type I IGF receptor (Fig. 3). We have addressed our attention to IGF-I receptor because this polypeptide affects on its own the synthesis of nuclear inositol lipids and bombesin seems only to synergize in eliciting the response. It is clear from Fig. 3 that type I IGF receptor is expressed in unresponsive cells to a higher extent as compared to responsive Swiss 3T3 fibroblasts. It seems that unresponsive cells attempt to override the impaired machinery of signal transduction inside the cell by overexpressing the receptor. The analysis of the phosphorylation pattern in both cell populations (Figs. 4 and 5) indicates that only responsive cells show a transient decrease of PIP and PIP 2 levels whilst no changes occur in nuclei purified from unresponsive Swiss 3T3 cells. This transient decrease is induced by IGF-I and bombesin only augments the effect since it is completely ineffective on its own. This raises the question of why these rapid changes occur at the nucleus after growth stimulation of quiescent cells by means of a polypeptide hormone (i.e., IGF-I) that uses a tyrosine kinase receptor at the plasma membrane and consequently a signalling system different from that elicited by polyphosphoinositide hydrolysis (19). It is worth pointing out that both in vitro phosphorylation of cytoplasmic fraction and in vivo labelling of whole cells (16, 17) clearly indicate that at the plasma membrane and/or at the cytoplasm IGF-I, after binding to its receptor, does not induce any changes in polyphosphoinositide phosphorylation nor any mass changes of these particular lipids. Therefore, it seems that two separate events occur at these early steps of induction of cell growth. The binding of IGF-I to its specific receptor induces the stimulation of tyrosine kinase activity at the plasma membrane and at the cytoplasm; after this, the nucleus responds using an autonomous signalling involving the metabolism of inositol lipids. This is supported by the evidence that only cells capable of replicating
FIG. 2. Electron microscopy of nuclei isolated from Swiss 3T3 cells as previously described (16). The inset shows details of the nuclear periphery. Note the absence of extranuclear debris and the complete stripping of the outer nuclear membrane. Bar = 1 pm.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION 1412tO
95
1251 I G F - I Binding Assay Specific Binding Expressed as C P M / 106ceLLs
8
6
2 0 Responsive c e L L s
Unresponsive ceiLs
FIG. 3. [1251] IGF-I binding assay in responsive and unresponsive Swiss 3T3 fibroblast. Cells were incubated at 4°C with labelled IGF-I (1 ng/ml) and unlabelled IGF-I (100 ng/ml) or unlabelled insulin (500/zg/ml). Control is constituted by the complete binding mixture with labelled IGF-I alone. This is repeated with excess of unlabelled IGF-I and insulin to test the specificity of binding and the receptor type.
their D N A under mitogenic stimulation show changes in PIP and PIP 2 phosphorylation levels in isolated nuclei and by experiments suggesting that the decrease of nuclear PIP and PIP 2 mass, assessed by in viva labelling with 3H-myo-inositol, are due to a PLC catalyzed hydrolysis of these lipids (17). Therefore, the availability of an unresponsive clone of Swiss 3T3 cells
1.8 1.6
7~ x
1.4
!!ii!i!i
1.2
[]P,P !iiiiii!IIP,P2
t.o g,
~
0.8
121 0.4
~°~i II 0.2
0
ilL i
lJ|
2
3
4
•
1 6
7
FIG. 4. Phosphorylation of polyphosphoinositides in nuclei from responsive Swiss 3T3 cells after mitogenic stimulation with IGF-I and/or bombesin. Mitogen concentrations as in Fig. 1. (1) control unstimulated cells; (2) bombesin 2 rain; (3) bombesin 1 hr; (4) IGF-I 2 min; (5) IGF-I 1 hr; (6) IGF-I plus bombesin 2 rain; (7) IGF-I plus bombesin 1 hr.
96
L. COCCO,et al. 20 1.8
q-
0
-x
1.6
1.4
•~
1.2
~_
I.O
PIP
I--
PIP2
~ 0,8 n
0.6
,.-.,
0.4 0.2
I
2
3
4
,5
6
7
FIG. 5. Phosphorylation of polyphosphoinositides in nuclei from unresponsive 3T3 cells after mitogenic stimulation. Conditions are as in Fig. 4.
was extremely important, once established that these cells express the type I IGF receptor, in order to ascertain whether the alterations in nuclear inositol lipids were likely to be causative. Indeed our data indicate that the onset of DNA synthesis takes place in responsive cells which, after binding IGF-I, undergo rapid changes in nuclear PIP and PIP 2 levels. On the contrary, in unresponsive cells something inside the cell is changed or unbalanced so that the metabolism of nuclear polyphosphoinositides is not affected. It seems that the cascade of events that begins at the plasma membrane does not reach the nucleus and in turn its signalling capability via inositoi lipid metabolites is absent. Thus, in the absence of a rapid hydrolysis of nuclear polyphosphoinositides, the onset of DNA synthesis does not take place. This raises the possibility that a crosstalk exists between the cell surface tyrosine kinase receptor for IGF-I and the metabolism of nuclear polyphosphoinositides. Moreover, when this communication is lacking, i.e., in unresponsive Swiss 3T3 cells, the rapid and transient hydrolysis of PIP and PIP 2 at the nucleus does not occur and subsequently the cells do not replicate. Therefore, it is possible to evidence a link between changes in inositol lipid metabolism and the onset of DNA synthesis. The presence at the nucleus of a pattern of newly synthesized DAG that differs from that of the cytoplasm in Friend cells (20) and the presence of IP 3 sensitive pools at the nuclear level (21) seem to address the question of whether these changes in polyphosphoinositides are also linked to changes in the localization of PKC whose activity has been previously shown to be present in nuclei of several cell types (22-27).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
97
T R A N S L O C A T I O N OF PKC TO THE NUCLEUS UNDER M I T O G E N I C S T I M U L A T I O N AND E F F E C T O F I N O S I T O L P H O S P H A T E S ON P K C A C T I V I T Y In the previous section we have focused on the issue of the existence of a peculiar nuclear phophoinositide signalling system hinted at by evidence that IGF-I, a tyrosine kinase stimulating growth factor, induces early and transient changes of nuclear PIP and PIP 2. Since the signal transduction system that utilizes the inositol lipids as precursors of second messengers, inositol phosphates and D A G , recognizes PKC as a target because of its physiological need for D A G and Ca 2+ to be activated (3), it seems relevant to investigate the fate of this phosphorylating enzyme after the same mitogenic stimulation which gives rise to changes in nuclear polyphosphoinositide levels. It has been shown that in quiescent Swiss 3T3 cells the in vivo phosphorylation of several nuclear proteins, namely 21, 31, 40, 45 and 66 kDa, is stimulated by IGF-I and bombesin within 45 rain of treatment (22). The in vitro assay using isolated nuclei from the same cells shows that the 31 kDa (i.e., histone H1) and the 21 kDa proteins are preferential substrates for PKC (22). By using the same experimental model employed for inositol lipid analysis, that is IGF-I responsive and unresponsive cells, it is possible to check the PKC activity in isolated nuclei after stimulation with IGF-I and bombesin. After 15 rain treatment the activity is evident and Figure 6 shows that only responsive 3T3 fibroblasts exhibit a marked increase of nuclear PKC
220 200 180
o
160
v
140
o
120
ceLLs []Responsive
I00
m UnresponsiveceLLs
80 60
o~
40 20 0 I
2
3
4.
FIG. 6. Nuclear PKC activityin vitro using exogenoushistone HI as substrate after treatment of responsive and unresponsive cells with IGF-I and bombesin. Values are averages of 5 separate experiments and refer to the percentage increase over the control (0). The relative amount of histone H1 band was determined by densitometric scanning of autoradiograms (18). (1) PS/DAG; (2) PS; (3) PS/DAG/SPH; (4) PS/DAG/EGTA.
98
L. COCCO, et al.
activity when compared to unresponsive fibroblasts. This clearly indicates that a stimulation of PKC dependent protein phosphorylation occurs only in nuclei of responsive cells and that this phenomenon follows temporally the changes of nuclear polyphosphoinositide metabolism. Nevertheless, a crucial question remains as to whether the mitogenic combination of IGF-I and bombesin activates PKC molecules already localized at nuclear level or induces PKC translocation from the cytoplasm to the nucleus. We have obtained an anti-PKC polyclonal antibody raised against the synthetic peptide CYVNPQVHPILQSAV derived from the C-terminal region of PKC and with a high sequence homology to alpha, beta, gamma and delta isoforms of the enzyme (28). The immunochemical analysis has been carried out only in the mitogen responsive cell population because of the absence of PKC activity in nuclei from unresponsive cells. Western blots of nuclear proteins (Fig. 7) clearly show that PKC transiocates to the nucleus after IGF-I and bombesin stimulation. In addition to a major staining band of 80 kDa corresponding to the native enzyme, the antibody recognizes also a second faint band of 50 kDa, presumably corresponding to the proteolytically derived, truncated PKC (PKM) since the antibody is specific for the catalytic domain (carboxy-terminal region) of the enzyme. To establish certainly that IGF-I and bombesin induce only translocation and not a de n o v o synthesis of PKC the immunoblot analysis has been extended to whole cell proteins from control and stimulated fibroblasts (Fig. 7). In this case the amount of total cellular PKC remains constant through the time course of the stimulation. Because of this evidence that PKC translocates to the nucleus when Swiss 3T3 cells are stimulated with IGF-I and bombesin and that in responsive cells this translocation follows temporally the changes in nuclear inositol lipid metabolism, as a step towards the elucidation of the role played by these lipids in eliciting the signals responsible for the onset of DNA synthesis, the effect of Ca 2+ free IP 3 on the PKC activity has been analyzed in nuclei from IGF-I and bombesin stimulated fibroblasts. When tested in vitro for its optimal activity PKC requires Ca 2÷, D A G and PS (22). In the absence of any exogenously added Ca 2+, PS and D A G are capable of stimulating the activity of nuclear PKC so that it is reasonable that the enzyme utilizes Ca 2+ ions already present in the nucleus of mitogen stimulated Swiss 3T3 cells (29). This suggestion is strengthened by observations dealing with the presence of Ca 2+ binding proteins in nuclei and suhnuclear fractions (30, 31). On the contrary, the presence of PS and D A G is essential because if only PS is added to the phosphorylation mixture, the PKC activity is almost negligible and this is also true for D A G (29). In the presence of IP 3 and without D A G the PKC activity in isolated nuclei is stimulated to the same levels obtained with the combination of PS and D A G (Fig. 8). This phenomenon is observable in a concentration range from 20 to 50 ~M.
1
2
3
4
1
2
3
4
80kDa,~
50kDa,.~
A
B
FIG. 7. Western blot analysis of Swiss 3T3 total cellular and nuclear proteins using an anti-PKC antibody (24). Panel A, nuclear proteins. Lane 1, partially purified PKC; lane 2, unstimulated cells; lane 3, stimulated cells for 45 min with IGF-I and bombesin; lane 4, as in lane 3 but incubated with rabbit preimmune serum. Panel B, total cellular proteins. Lane 1, quiescent cells; lane 2, cells stimulated for 5 min with IGF-I and bombesin; lane 3, cells stimulated for 15 min with IGF-I and bombesin; lane 4, cells stimulated for 45 rain with IGF-I and bombesin.
FIG. 8. In vitro phosphorylation of purified nuclei from 45 min mitogen-stimulated 3T3 cells in the presence of histone H1 as exogenous substrate for PKC activity. Reactions were carried out in the absence of exogenous Ca 2+. Left panel is the Coomassie Blue staining. Right panel is the corresponding autoradiogram. The first lane on the left corresponds to molecular weight standards.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
99
220 200 180 160 140 120
~=
15rain. treatment
I00
I=
45min.treatment
80 60 40 20 O' ]
2
3
4
5
6
7
FIG. 9. PKC activity in the presence of IPa and PKC inhibitors in isolated nuclei of Swiss 3T3 cells stimulated with IGF-I and bombesin for 15 or 45 min. Histone H1 was used as exogenous substrate. Relative amount of radioactivity in the H1 band was determined by densitometric scanning of the autoradiograms. Values are the average of 5 separate experiments and refer to percentage increase over the control (0). (1) PS/DAG; (2) PS; (3) PS/DAG/SPH; (4) PS/DAG/EGTA; (5) PS/IP3; (6) PS/IP3/EGTA; (7) PS/IP3/SPH.
Figure 8 shows also that all of the other inositol phosphates tested together with PS are ineffective. The IP 3 dependent stimulation of nuclear P K C is abolished by either the P K C inhibitor sphingosine or by preincubating the nuclei with E G T A (Fig. 9). The same Figure shows that the P K C stimulation by IP 3 occurs during the time course of the mitogenic stimulation. The ability of E G T A to abolish the IP 3 induced stimulation of nuclear P K C suggests a direct involvement of Ca 2+. Because of the discovery of IP 3 receptors at the nuclear envelope (32) and the temporal relationship between changes in nuclear inositol lipids and P K C translocation to the nucleus it seems possible that second messengers such as IP 3 are produced and can act as a stimulatory molecule for P K C and the above reported in vitro analysis supports this possibility. It is well known that D A G lowers the threshold of P K C for Ca2÷; therefore, it is conceivable that IP 3 mimics the role of D A G by liberating Ca 2+ from nuclear stores. CONCLUSIONS The occurrence of lipids in the cell nucleus has been hinted at by several observations (9-11). Nevertheless, this review does not deal with the localization and structural functions of the more abundant phospholipids but focuses attention on the existence of an a u t o n o m o u s nuclear signalling system that utilizes polyphosphoinositides which are peculiar indeed a m o n g other phospholipids (1).
100
L. COCCO, et al.
The growth promoting effect of IGF-I begins at the plasma membrane with the binding to its tyrosine kinase receptor and the downstream stage of the cascade seems to be constituted by the rapid hydrolysis of nuclear polyphosphoinositides. One might argue that the mitogenic stimulation described is given by a combination of IGF-I and bombesin, the latter is capable of activating at the plasma membrane the specific PLC for PIP 2 (33). However, it is clear from the analysis of the replicative capability after mitogenic stimulation that IGF-I on its own induces DNA synthesis and bombesin acts as a synergizing agent. Moreover, only IGF-I affects both nuclear inositol lipids and PKC activity whereas bombesin is completely ineffective. The changes produced at nuclear level by IGF-I are visible only in Swiss 3T3 cells which replicate after mitogenic stimulation. Once established that the type I IGF receptor was expressed at the plasma membrane in both responsive and unresponsive cells the early changes in nuclear polyphosphoinositide levels appear to be actually causative in eliciting the onset of DNA synthesis. It is rather intriguing to establish all the steps by which the signal generated by IGF-I stimulation reaches the nucleus; however, it is well accepted that the basic operating principle is that an extracellular signal affects the activity of a protein kinase cascade and this is certainly the case for IGF-I which stimulates tyrosine kinase activity (19). Therefore, it is clear that when IGF-I binds to its receptor a phosphorylation cascade takes place and this signal reaches the nucleus. At this stage a rapid and transient hydrolysis of nuclear PIP and PIP 2 occurs and this precedes the translocation of PKC to the nucleus. The fact that this step is causative for the onset of DNA synthesis is hinted at by the evidence that only in responsive cells do nuclear polyphosphoinositide changes and PKC translocation occur. When the same mitogenic stimulus (i.e., IGF-I) is applied to unresponsive cells, which still express the IGF-I tyrosine kinase receptor, something inside the cell is imbalanced and the crosstalk between surface receptor and nuclear compartment is lacking, where the signalling system is not elicited and then DNA synthesis does not take place. The in v i t r o stimulation of nuclear PKC by means of Ca 2+ free IP 3 supports the direct link between inositol lipid hydrolysis and PKC activation. Moreover, the PLC activated hydrolysis of PIP and PIP 2 also generates DAG, whose levels at the nucleus undergo changes during cell differentiation (20) and during mitogenic stimulation (unpublished observations), so that it is conceivable that D A G can lower the threshold of nuclear PKC for Ca 2+. In addition, the PLC activated hydrolysis of PIP produces IP 2 which has been suggested to have an intranuclear function in directly stimulating DNA polymerase activity (34). Therefore, we can conclude that among IGF-I stimulated events in nuclei of Swiss 3T3 cells the changes of inositol lipids and the translocation of PKC represent a key step for the subsequent onset of DNA synthesis.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
101
SUMMARY T h e correlation between changes in nuclear polyphosphoinositide levels preceding P K C translocation to the nucleus and the onset of D N A synthesis has been discussed. Using two different clones of Swiss 3T3 fibroblasts belonging to the same original cell line, one of which is unresponsive to mitogenic stimulation with I G F - I on its own or in combination with bombesin, it has been observed that a rapid and transient breakdown of nuclear PIP and PIP E occurs only in responsive cells and this precedes the translocation of P K C to the nucleus, as evidenced by immunochemical analysis as well as by enzymatic activity. Therefore, it seems that a direct link exists between nuclear polyphosphoinositide metabolism, P K C translocation to the nucleus and cell division. Since I G F - I acts at the plasma m e m b r a n e through a tyrosine kinase receptor it seems that the mitogenic stimulation induced by this factor utilizes different signalling pathways at the plasma m e m b r a n e and at the nucleus. Because of the evidence that type I I G F receptor is expressed in both responsive and unresponsive cells and that the receptor machinery at the plasma m e m b r a n e is active the lack of the transient changes in nuclear inositol lipids and of P K C translocation in unresponsive cells further suggests that the cell nucleus is capable of an a u t o n o m o u s signalling system based on polyphosphoinositide metabolism. ACKNOWLEDGEMENTS The authors wish to express their appreciation to Dr. S. Capitani for his helpful advice and c o m m e n t s , to Dr. E. Faicieri for electron microscopy observations and to A. Fantazzini for his skilled technical assistance. This work was supported by Italian C . N . R . grants PF I G and PF BTBS.
REFERENCES 1. M.J. BERR1DGE and R. F. IRVINE, lnositol trisphosphate, a novel second messenger in cellular signal transduction, Nature 312,315-321 (1984). 2. C.P. DOWNES and C. H. MACPHEE, Myo-inositol metabolites and cellular signals, Eur. J. Biochem. 193, 1-18 (1990). 3. Y. NISHIZUKA, The molecular heterogeneity of protein kinase C and its implications for cellular regulation, Nature 334, 661-665 (1988). 4. M.J. BERRIDGE and R. F. IRVINE, lnositol phosphates and cell signalling, Nature 341,197-204 (1989). 5. E. ROZENGURT, Signal transduction pathways in mitogenesis. Brit. Med. Bull. 45, 515-528 (1989). 6. R. M. EVANS, The steroid and thyroid hormone receptor family, Science 240, 889-895 (1988). 7. L. COCCO, R. S. GILMOUR, A OGNIBENE, A. J. LETCHER, F. A. MANZOLI and R. F. IRVINE, Synthesis of polyphosphoinositides in nuclei of Friend cells.
102
8. 9. 10. 11. 12. 13. 14.
15.
16.
17. 18.
19. 20. 21. 22. 23. 24.
L. COCCO, et al. Evidence for polyphosphoinositide metabolism inside the nucleus which changes with cell differentiation, Biochern. J. 248, 765-770 (1987). F. A. MANZOLI, A. M. MARTELLI, S. CAPITANI, N. M. MARALDI, R. RIZZOLI, O. B A R N A B E I and L. COCCO, Nuclear polyphosphoinositides during cell growth and differentiation, Advan. Enzyme Regul. 28, 25-34 (1989). F. A. MANZOLI, S. CAPITANI and N. M. MARALDI, Minor components of chromatin and their role in the release of template restriction, pp. 463--486 in Cell Growth (C. NICOLINI, ed.), Plenum Press, New York (1982). F.A. MANZOLI, N. M. MARALDI, S. CAPITANI, L. COCCO and O. BARNABEI, Chromatin lipids and their possible role in gene expression. A study in normal and neoplastic cells, Advan. Enzyme Regul. 17, 175-184 (1979). F . A . MANZOLI, N. M. MARALDI, L. COCCO, S. CAPITANI and A. FACCHINI, Chromatin phospholipids in normal and chronic lymphocytic leukemia lymphocytes, Cancer Res. 37, 843-849 (1977). S. CAPITANI, L. COCCO, N. M. MARALDI, S. PAPA and F. A. MANZOLI, Effects of phospholipids on transcription and ribonucleoprotein processing in isolated nuclei, Advan. Enzyme Regul. 25, 425-438 (1986). L. COCCO, N. M. MARALDI, F. A. MANZOLI, R. S. GILMOUR and A. LANG, Phospholipid interactions in rat liver nuclear matrix, Biochern. Biophys. Res. Commun. 96, 890-898 (1980). S. CAPITANI, V. BERTAGNOLO, M. MAZZONI, P. SANTI, M. PREVIATI, A. ANTONUCCI and F. A. MANZOLI, Lipid phosphorylation in isolated rat liver nuclei. Synthesis of polyphosphoinositides at subnuclear level, FEBS Lett. 254, 194-198 (1989). S. CAPITANI, B. HELMS, M. MAZZONI, M. PREVIATI, V. BERTAGNOLO, K. W. A. W l R T Z and F. A. MANZOLI, Uptake and phosphorylation of phosphatidylinositol by rat liver nuclei. Role of phosphatidylinositol transfer protein, Biochim. Biophys. Acta 1044, 193-200 (1990). L. COCCO, A. M. MARTELLI, R. S. GILMOUR, A. OGNIBENE, F. A. MANZOLI and R. F. IRVINE, Rapid changes in phospholipid metabolism within nuclei of Swiss 3T3 cells induced by treatment of the cells with insulin-like growth factor I, Biochern. Biophys. Res. Commun. 154, 1266-1272 (1988). L. COCCO, A. M. MARTELLI, R. S. GILMOUR, A. OGNIBENE, F. A. MANZOLI and R. F. IRVINE, Changes in nuclear inositol phospholipids induced in intact cells by insulin-like growth factor I, Biochem. Biophys. Res. Comrnun. 159, 720-725 (1989). A . M . MARTELLI, R. S. GILMOUR, L. M. NERI, L. MANZOLI, A. N. CORPS and L. COCCO, Mitogen-stimulated events in nuclei of Swiss 3T3 cells. Evidence for a direct link between changes of inositol lipids, protein kinase C requirement and the onset of DNA synthesis, FEBS Lett. 283, 243-246 (1991). R . S . GILMOUR, C. G. PROSSER, I. R. FLEET, L. COCCO, J. C. SAUNDERS, K. D. BROWN and A. N. CORPS, From animal to molecule: aspects of the biology of insulin-like growth factor I, Brit. J. Cancer 58, 23-30 (1988). O. TRUBIANI, A. M. MARTELLI, L. MANZOLI, E. SANTAVENERE and L. COCCO, Nuclear lipids in Friend cells. Shifted profile of diacylglycerol during erythroid differentiation induced by DMSO, Cell Biol. Intl. Rep. 14, 559-566 (1990). P. NICOTERA, S. ORRENIUS, T. NILSSON and P.-O. BERGGREN, An inositol 1,4,5-trisphosphate-sensitive Ca 2÷ pool in liver nuclei, Proc. Natl. Acad. Sci. U.S.A. 87, 6858-6862 (1990). A. M. MARTELLI, R. S. GILMOUR, E. FALC1ERI, F. A. MANZOLI and L. COCCO, Mitogen-stimulated phosphorylation of nuclear proteins in Swiss 3T3 cells: evidence for a protein kinase C requirement, Exptl. Cell Res. 185, 191-202 (1989). T. P. THOMAS, H. S. T A L W A R and W. B. ANDERSON, Phorbol ester-mediated association of protein kinase C to the nuclear fraction in NIH 3T3 cells, Cancer Res. 48, 1910-1919 (1988). A . R . BUCKLEY, P. D. CROWE and D. HADDOCK-RUSSELL, Rapid activation of protein kinase C in isolated rat liver nuclei by prolactin, a known hepatic mitogen, Proc. Natl. Acad. Sci. U.S.A. 85, 8649-8653 (1988).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION 25. 26. 27.
28.
29.
30. 31. 32. 33. 34.
103
A. P. FIELDS, G. TYLER, A. S. KRAFT and W. S. STRADFORD MAY, Role of nuclear protein kinase C in the mitogenic response to platelet-derived growth factor, J. Cell Sci. 96, 107-114 (1990). S. CAPITANI, P. R. GIRARD, G. J. MAZZEI, J. F. KUO, R. BEREZNEY and F. A. MANZOLI, Immunochemical characterization of protein kinase C in rat liver nuclei and subnuclear fractions, Biochem. Biophys. Res. Commun. 142, 367-375 (1987). A . M . MARTELLI, C. CARIN1, S. MARMIROLI, M. MAZZONI, P. J. BARKER, R. S. GILMOUR and S. CAPITANI, Nuclear protein kinases in rat liver: evidence for increased histone H1 phosphorylating activity during liver regeneration, Exptl. Cell Res. 195, 255--262 (1991). A . M . MARTELLI, R. S. GILMOUR, F. A. MANZOLI and L. COCCO, Calcium free inositol (1,4,5)-trisphosphate stimulates protein kinase C dependent protein phosphorylation in nuclei isolated from mitogen-treated Swiss 3T3 cells, Biochem. Biophys. Res. Commun. 173, 149-155 (1990). A. M. MARTELLI, L. M. NERI, R. S. GILMOUR, P. J. BARKER, N. S. HUSKISSON, F. A. MANZOLI and L. COCCO, Temporal changes in intracellular distribution of protein kinase C in Swiss 3T3 cells during mitogenic stimulation with insulin-like growth factor I and bombesin: translocation to the nucleus follows rapid changes in nuclear polyphosphoinositides, Biochem. Biophys. Res. Commun. 177, 480--487 (1991). O. BACHS and E. CARAFOLI, Calmodulin and calmodulin-binding proteins in liver cell nuclei, J. Biol. Chem. 262, 10786--10790 (1987). M.J. PUJOL, M. SORIANO, R. ALIGUE', E. CARAFOLI and O. BACHS, Effect of alpha-adrenergic blockers on calmodulin association with the nuclear matrix of rat liver cells during proliferative activation, J. Biol. Chem. 264, 18863-18865 (1989). C . A . ROSS, J. MELDOLESI, T. A. MILNER, T. SATOH, S. SUPAq~I'APONE and S. SNYDER, Inositol 1,4,5-trisphosphate receptor localized to endoplasmic reticulum in cerebellar Purkinije neurons, Nature 339, 468--470 (1989). J. P. HESLOP, D. M. BLAKELEY, K. D. BROWN, R. F. IRV1NE and M. J. BERRIDGE, Effects of bombesin and insulin on inositol (1,4,5) trisphosphate and inositol (1,3,4) trisphosphate formation in Swiss 3T3 cells, Cell 47, 703-709 (1986). V. SYLVIA, G. CURTIN, J. NORMAN, J. STEC and D. BUSBEE, Activation of a low specific activity form of DNA polymerase alpha by inositol-l,4-bisphosphate, Cell 54, 651~558 (1988).
0065-2571/92/$15.00 © 1992Pergamon Press pie.
Printed in Great Britain. All rights reserved
CHANGES IN INOSITOL LIPID METABOLISM AND PROTEIN KINASE C TRANSLOCATION IN NUCLEI OFMITOGEN STIMULATED SWISS 3T3 CELLS LUCIO COCCO*, ALBERTO M. MARTELLI*, R. STEWART GILMOUR§, ROSA ALBA RANAt, OT'I'AVIO BARNABEI:~ and FRANCESCO A. MANZOLI* *Institute of Human Anatomy, University of Bologna, Italy tlnstitute of Human Anatomy, University of Chieti, Italy :[:Department of Biology, University of Bologna, Italy §Department of Molecular and Cellular Physiology, Institute of Animal Physiology and Genetics Research, AFRC, Babraham, Cambridge, UK INTRODUCTION
The regulation of genetic material must be intimately coordinated with cellular as well as extracellular events. Clearly the cell is capable of transducing information from the periphery into the nucleus. An understanding of how this information is transmitted requires a thorough knowledge of both the structural organization of the nucleus and the metabolic pathway of the signalling system. Inositol lipid metabolism has become closely identified with the production of multiple informational molecules (1) and the awareness that a phospholipid can act as signal reservoir has become increasingly accepted (2). Indeed the identification of DAG as a physiological regulator of PKC has provided the first good evidence that lipids themselves are actively involved in signal transduction (3). The interest has been augmented by the evidence that IP3, generated by the PLC specific for PIP2, is able to mobilize calcium ions from intracellular stores; thus, a key role for lipid second messengers, inositol phosphates, is to modulate the fluxes inside the cell of the ubiquitous second messenger, Ca 2+ (4). During mitogenic stimulation a number of events takes place. First of all, there is the specific binding of a growth factor to its own receptor on the plasma membrane. Information dealing with the following early Abbreviations: IP3, inositol-l,4,5-phosphate; DAG, diacylglycerol; ATP, adenosine triphosphate; PI, phosphatidylinositol; cAMP, cyclic-adenosine monophosphate; IGF-I, insulin-like growth factor-I; PIP, phosphatidylinositol-4-phosphate; PIP2, phosphatidylinositol-4,5-phosphate; PA, phosphatidic acid; BrdU, 5-bromodeoxyuridine; PKC, protein kinase C; PS, phosphatidylserine; EGTA, (ethylenedinitrilo) tetraacetic acid; SPH, sphingosine; PLC, phospholipase C. 91
92
L. COCCO,et al.
events indicates that some polypeptide hormones, acting as growth factors, induce a transient appearance of cytoplasmic second messengers such as IP 3 and DAG after binding to their own receptor, whereas other factors, acting through a tyrosine kinase receptor, directly activate protein phosphorylation and finally other hormones use the cAMP as second messenger (5). Therefore, the signalling pathways elicited by cell surface receptors activated by growth factors are more complicated as compared to the signals that activate nuclear receptors. These signals indeed are transmitted by small amphiphilic molecules such as steroids that can freely diffuse through lipid membranes and reach the nucleus (6). Nevertheless, the mechanism by which the mitogenic stimulus generated by polypeptide growth factors is eventually transmitted to the nucleus and whether or not a signalling system exists at the nucleus itself remains almost unknown. The evidence for the occurrence at nuclear level of inositol lipid metabolism (7) and the relationship between this metabolism and cell growth and differentiation (8) has suggested that inside the nucleus a signalling system could exist using polyphosphoinositides as messengers. NUCLEAR INOSITOL LIPIDS AND MITOGENIC STIMULATION WITH IGF-I AND BOMBESIN IN SWISS 3T3 CELLS Phospholipids have been identified in the cell nucleus and their pattern differs from that of the whole cell (9). The evidence that they are components of the chromatin adds to their obvious presence in the outer nuclear membrane (10) and it has been reported that changes occur in the nuclear phospholipids during neoplastic transformation (11). Their possible action has been analyzed in vitro and it has been shown that they can modulate replication and transcription and mediate the attachment of newly synthesized DNA to the nuclear matrix (12, 13). Moreover, it has been shown that isolated nuclei are able to synthesize PA, PIP and PIP 2 and that after induction of differentiation of Friend cells the pattern of inositoi lipid phosphorylation changes at nuclear level in a manner consistent with a decrease in PIP phosphorylation and an increase in PIP 2 synthesis (7). Certainly these pioneering observations pointed to the likelihood that nature has put these unique lipids to some important use inside the nucleus in addition to their functions outside it. Interestingly, it has been shown that these phosphorylative events take place also independently from the nuclear envelope, since the phosphorylation of inositol lipids occurs in nuclear matrix and in purified nuclear lamina (14). In addition, isolated nuclei have been shown to incorporate and phosphorylate exogenous PI delivered by a specific PI-transfer protein (15).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
93
Recently we have shown that isolated nuclei of Swiss 3T3 fibroblasts incorporate ATP into inositol lipids and that changes occur in their metabolism when quiescent cells are stimulated to grow with a mitogenic combination of IGF-I and bombesin. Actually this effect on inositol lipid phosphorylation is mainly due to IGF-I which induces a decrease of incorporation of [T32p] ATP into PIP and PIP 2 whereas bombesin on its own has no effect. The changes are transient in that, although they are still detectable after 15 min, after 1 hr they are no longer significant (16). During routine passages of cells we selected a clone still maintaining contact inhibition but unresponsive to the above mitogens. This provided an opportunity to investigate a possible correlation between the onset of DNA synthesis and the early changes in nuclear polyphosphoinositides previously described (17). The capability to respond to growth factor stimulation has been evaluated by analyzing DNA synthesis by means of BrdU incorporation (18). Only responsive cells after mitogenic stimulation incorporate BrdU as demonstrated by immunofluorescence microscopy. The number of responsive cells which replicate their DNA is directly related to the stimulus given (Fig. 1). IGF-I in combination with bombesin is the more efficient condition, while both IGF-I and bombesin on their own give rise to a lower number of replicating cells. The unresponsive cells do not incorporate BrdU either in the presence of IGF-I and bombesin on their own or in the presence of the mitogenic combination.
80I
Responsive ceLLs • Unresponsive ceLLs
70
60
"0
50
o 40 30 2O l0 m
I
2
3
4
FIG. 1. Effect of IGF-| and/or bombesin on stimulation of DNA synthesis in Swiss 3'13 cells. DNA synthesis was assessed by BrdU incorporation after growth stimulation of quiescent cells with the above growth factors for 20 hr. Cells were labelled for 60 rain with 10 p~MBrdU and 1 ~M fluorodeoxyuridine and then processed for immunofluorescence microscopy (18). The histogram represents the percentage of labelled nuclei. (1) control unstimulated cells; (2) IGF-I 20 ng/ml; (3) bombesin 1 nM; (4) IGF-I plus bombesin.
94
L. COCCO, et al.
Three points are essential for establishing a correlation between nuclear inositol lipids and the onset of DNA synthesis: (i) the purity of nuclear preparation; (ii) the expression of type I IGF receptor in unresponsive cells; (iii) the lack of changes in polyphosphoinositide metabolism in unresponsive cells. Figure 2 shows that, as assessed by electron microscopy analysis on a large number of thin sections, nuclei purified from Swiss 3T3 cells in the presence of non-ionic detergent (16) are highly pure and lack the outer membrane and extranuclear debris. The cytoplasmic enzymatic markers are indeed absent since in nuclear preparation we were dealing with the activity of glucose-6-phosphatase is less than 0.1% of the activity present in whole cell homogenates. Swiss 3T3 fibroblasts which do not replicate under mitogenic stimulation as well as responsive cells express the type I IGF receptor (Fig. 3). We have addressed our attention to IGF-I receptor because this polypeptide affects on its own the synthesis of nuclear inositol lipids and bombesin seems only to synergize in eliciting the response. It is clear from Fig. 3 that type I IGF receptor is expressed in unresponsive cells to a higher extent as compared to responsive Swiss 3T3 fibroblasts. It seems that unresponsive cells attempt to override the impaired machinery of signal transduction inside the cell by overexpressing the receptor. The analysis of the phosphorylation pattern in both cell populations (Figs. 4 and 5) indicates that only responsive cells show a transient decrease of PIP and PIP 2 levels whilst no changes occur in nuclei purified from unresponsive Swiss 3T3 cells. This transient decrease is induced by IGF-I and bombesin only augments the effect since it is completely ineffective on its own. This raises the question of why these rapid changes occur at the nucleus after growth stimulation of quiescent cells by means of a polypeptide hormone (i.e., IGF-I) that uses a tyrosine kinase receptor at the plasma membrane and consequently a signalling system different from that elicited by polyphosphoinositide hydrolysis (19). It is worth pointing out that both in vitro phosphorylation of cytoplasmic fraction and in vivo labelling of whole cells (16, 17) clearly indicate that at the plasma membrane and/or at the cytoplasm IGF-I, after binding to its receptor, does not induce any changes in polyphosphoinositide phosphorylation nor any mass changes of these particular lipids. Therefore, it seems that two separate events occur at these early steps of induction of cell growth. The binding of IGF-I to its specific receptor induces the stimulation of tyrosine kinase activity at the plasma membrane and at the cytoplasm; after this, the nucleus responds using an autonomous signalling involving the metabolism of inositol lipids. This is supported by the evidence that only cells capable of replicating
FIG. 2. Electron microscopy of nuclei isolated from Swiss 3T3 cells as previously described (16). The inset shows details of the nuclear periphery. Note the absence of extranuclear debris and the complete stripping of the outer nuclear membrane. Bar = 1 pm.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION 1412tO
95
1251 I G F - I Binding Assay Specific Binding Expressed as C P M / 106ceLLs
8
6
2 0 Responsive c e L L s
Unresponsive ceiLs
FIG. 3. [1251] IGF-I binding assay in responsive and unresponsive Swiss 3T3 fibroblast. Cells were incubated at 4°C with labelled IGF-I (1 ng/ml) and unlabelled IGF-I (100 ng/ml) or unlabelled insulin (500/zg/ml). Control is constituted by the complete binding mixture with labelled IGF-I alone. This is repeated with excess of unlabelled IGF-I and insulin to test the specificity of binding and the receptor type.
their D N A under mitogenic stimulation show changes in PIP and PIP 2 phosphorylation levels in isolated nuclei and by experiments suggesting that the decrease of nuclear PIP and PIP 2 mass, assessed by in viva labelling with 3H-myo-inositol, are due to a PLC catalyzed hydrolysis of these lipids (17). Therefore, the availability of an unresponsive clone of Swiss 3T3 cells
1.8 1.6
7~ x
1.4
!!ii!i!i
1.2
[]P,P !iiiiii!IIP,P2
t.o g,
~
0.8
121 0.4
~°~i II 0.2
0
ilL i
lJ|
2
3
4
•
1 6
7
FIG. 4. Phosphorylation of polyphosphoinositides in nuclei from responsive Swiss 3T3 cells after mitogenic stimulation with IGF-I and/or bombesin. Mitogen concentrations as in Fig. 1. (1) control unstimulated cells; (2) bombesin 2 rain; (3) bombesin 1 hr; (4) IGF-I 2 min; (5) IGF-I 1 hr; (6) IGF-I plus bombesin 2 rain; (7) IGF-I plus bombesin 1 hr.
96
L. COCCO,et al. 20 1.8
q-
0
-x
1.6
1.4
•~
1.2
~_
I.O
PIP
I--
PIP2
~ 0,8 n
0.6
,.-.,
0.4 0.2
I
2
3
4
,5
6
7
FIG. 5. Phosphorylation of polyphosphoinositides in nuclei from unresponsive 3T3 cells after mitogenic stimulation. Conditions are as in Fig. 4.
was extremely important, once established that these cells express the type I IGF receptor, in order to ascertain whether the alterations in nuclear inositol lipids were likely to be causative. Indeed our data indicate that the onset of DNA synthesis takes place in responsive cells which, after binding IGF-I, undergo rapid changes in nuclear PIP and PIP 2 levels. On the contrary, in unresponsive cells something inside the cell is changed or unbalanced so that the metabolism of nuclear polyphosphoinositides is not affected. It seems that the cascade of events that begins at the plasma membrane does not reach the nucleus and in turn its signalling capability via inositoi lipid metabolites is absent. Thus, in the absence of a rapid hydrolysis of nuclear polyphosphoinositides, the onset of DNA synthesis does not take place. This raises the possibility that a crosstalk exists between the cell surface tyrosine kinase receptor for IGF-I and the metabolism of nuclear polyphosphoinositides. Moreover, when this communication is lacking, i.e., in unresponsive Swiss 3T3 cells, the rapid and transient hydrolysis of PIP and PIP 2 at the nucleus does not occur and subsequently the cells do not replicate. Therefore, it is possible to evidence a link between changes in inositol lipid metabolism and the onset of DNA synthesis. The presence at the nucleus of a pattern of newly synthesized DAG that differs from that of the cytoplasm in Friend cells (20) and the presence of IP 3 sensitive pools at the nuclear level (21) seem to address the question of whether these changes in polyphosphoinositides are also linked to changes in the localization of PKC whose activity has been previously shown to be present in nuclei of several cell types (22-27).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
97
T R A N S L O C A T I O N OF PKC TO THE NUCLEUS UNDER M I T O G E N I C S T I M U L A T I O N AND E F F E C T O F I N O S I T O L P H O S P H A T E S ON P K C A C T I V I T Y In the previous section we have focused on the issue of the existence of a peculiar nuclear phophoinositide signalling system hinted at by evidence that IGF-I, a tyrosine kinase stimulating growth factor, induces early and transient changes of nuclear PIP and PIP 2. Since the signal transduction system that utilizes the inositol lipids as precursors of second messengers, inositol phosphates and D A G , recognizes PKC as a target because of its physiological need for D A G and Ca 2+ to be activated (3), it seems relevant to investigate the fate of this phosphorylating enzyme after the same mitogenic stimulation which gives rise to changes in nuclear polyphosphoinositide levels. It has been shown that in quiescent Swiss 3T3 cells the in vivo phosphorylation of several nuclear proteins, namely 21, 31, 40, 45 and 66 kDa, is stimulated by IGF-I and bombesin within 45 rain of treatment (22). The in vitro assay using isolated nuclei from the same cells shows that the 31 kDa (i.e., histone H1) and the 21 kDa proteins are preferential substrates for PKC (22). By using the same experimental model employed for inositol lipid analysis, that is IGF-I responsive and unresponsive cells, it is possible to check the PKC activity in isolated nuclei after stimulation with IGF-I and bombesin. After 15 rain treatment the activity is evident and Figure 6 shows that only responsive 3T3 fibroblasts exhibit a marked increase of nuclear PKC
220 200 180
o
160
v
140
o
120
ceLLs []Responsive
I00
m UnresponsiveceLLs
80 60
o~
40 20 0 I
2
3
4.
FIG. 6. Nuclear PKC activityin vitro using exogenoushistone HI as substrate after treatment of responsive and unresponsive cells with IGF-I and bombesin. Values are averages of 5 separate experiments and refer to the percentage increase over the control (0). The relative amount of histone H1 band was determined by densitometric scanning of autoradiograms (18). (1) PS/DAG; (2) PS; (3) PS/DAG/SPH; (4) PS/DAG/EGTA.
98
L. COCCO, et al.
activity when compared to unresponsive fibroblasts. This clearly indicates that a stimulation of PKC dependent protein phosphorylation occurs only in nuclei of responsive cells and that this phenomenon follows temporally the changes of nuclear polyphosphoinositide metabolism. Nevertheless, a crucial question remains as to whether the mitogenic combination of IGF-I and bombesin activates PKC molecules already localized at nuclear level or induces PKC translocation from the cytoplasm to the nucleus. We have obtained an anti-PKC polyclonal antibody raised against the synthetic peptide CYVNPQVHPILQSAV derived from the C-terminal region of PKC and with a high sequence homology to alpha, beta, gamma and delta isoforms of the enzyme (28). The immunochemical analysis has been carried out only in the mitogen responsive cell population because of the absence of PKC activity in nuclei from unresponsive cells. Western blots of nuclear proteins (Fig. 7) clearly show that PKC transiocates to the nucleus after IGF-I and bombesin stimulation. In addition to a major staining band of 80 kDa corresponding to the native enzyme, the antibody recognizes also a second faint band of 50 kDa, presumably corresponding to the proteolytically derived, truncated PKC (PKM) since the antibody is specific for the catalytic domain (carboxy-terminal region) of the enzyme. To establish certainly that IGF-I and bombesin induce only translocation and not a de n o v o synthesis of PKC the immunoblot analysis has been extended to whole cell proteins from control and stimulated fibroblasts (Fig. 7). In this case the amount of total cellular PKC remains constant through the time course of the stimulation. Because of this evidence that PKC translocates to the nucleus when Swiss 3T3 cells are stimulated with IGF-I and bombesin and that in responsive cells this translocation follows temporally the changes in nuclear inositol lipid metabolism, as a step towards the elucidation of the role played by these lipids in eliciting the signals responsible for the onset of DNA synthesis, the effect of Ca 2+ free IP 3 on the PKC activity has been analyzed in nuclei from IGF-I and bombesin stimulated fibroblasts. When tested in vitro for its optimal activity PKC requires Ca 2÷, D A G and PS (22). In the absence of any exogenously added Ca 2+, PS and D A G are capable of stimulating the activity of nuclear PKC so that it is reasonable that the enzyme utilizes Ca 2+ ions already present in the nucleus of mitogen stimulated Swiss 3T3 cells (29). This suggestion is strengthened by observations dealing with the presence of Ca 2+ binding proteins in nuclei and suhnuclear fractions (30, 31). On the contrary, the presence of PS and D A G is essential because if only PS is added to the phosphorylation mixture, the PKC activity is almost negligible and this is also true for D A G (29). In the presence of IP 3 and without D A G the PKC activity in isolated nuclei is stimulated to the same levels obtained with the combination of PS and D A G (Fig. 8). This phenomenon is observable in a concentration range from 20 to 50 ~M.
1
2
3
4
1
2
3
4
80kDa,~
50kDa,.~
A
B
FIG. 7. Western blot analysis of Swiss 3T3 total cellular and nuclear proteins using an anti-PKC antibody (24). Panel A, nuclear proteins. Lane 1, partially purified PKC; lane 2, unstimulated cells; lane 3, stimulated cells for 45 min with IGF-I and bombesin; lane 4, as in lane 3 but incubated with rabbit preimmune serum. Panel B, total cellular proteins. Lane 1, quiescent cells; lane 2, cells stimulated for 5 min with IGF-I and bombesin; lane 3, cells stimulated for 15 min with IGF-I and bombesin; lane 4, cells stimulated for 45 rain with IGF-I and bombesin.
FIG. 8. In vitro phosphorylation of purified nuclei from 45 min mitogen-stimulated 3T3 cells in the presence of histone H1 as exogenous substrate for PKC activity. Reactions were carried out in the absence of exogenous Ca 2+. Left panel is the Coomassie Blue staining. Right panel is the corresponding autoradiogram. The first lane on the left corresponds to molecular weight standards.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
99
220 200 180 160 140 120
~=
15rain. treatment
I00
I=
45min.treatment
80 60 40 20 O' ]
2
3
4
5
6
7
FIG. 9. PKC activity in the presence of IPa and PKC inhibitors in isolated nuclei of Swiss 3T3 cells stimulated with IGF-I and bombesin for 15 or 45 min. Histone H1 was used as exogenous substrate. Relative amount of radioactivity in the H1 band was determined by densitometric scanning of the autoradiograms. Values are the average of 5 separate experiments and refer to percentage increase over the control (0). (1) PS/DAG; (2) PS; (3) PS/DAG/SPH; (4) PS/DAG/EGTA; (5) PS/IP3; (6) PS/IP3/EGTA; (7) PS/IP3/SPH.
Figure 8 shows also that all of the other inositol phosphates tested together with PS are ineffective. The IP 3 dependent stimulation of nuclear P K C is abolished by either the P K C inhibitor sphingosine or by preincubating the nuclei with E G T A (Fig. 9). The same Figure shows that the P K C stimulation by IP 3 occurs during the time course of the mitogenic stimulation. The ability of E G T A to abolish the IP 3 induced stimulation of nuclear P K C suggests a direct involvement of Ca 2+. Because of the discovery of IP 3 receptors at the nuclear envelope (32) and the temporal relationship between changes in nuclear inositol lipids and P K C translocation to the nucleus it seems possible that second messengers such as IP 3 are produced and can act as a stimulatory molecule for P K C and the above reported in vitro analysis supports this possibility. It is well known that D A G lowers the threshold of P K C for Ca2÷; therefore, it is conceivable that IP 3 mimics the role of D A G by liberating Ca 2+ from nuclear stores. CONCLUSIONS The occurrence of lipids in the cell nucleus has been hinted at by several observations (9-11). Nevertheless, this review does not deal with the localization and structural functions of the more abundant phospholipids but focuses attention on the existence of an a u t o n o m o u s nuclear signalling system that utilizes polyphosphoinositides which are peculiar indeed a m o n g other phospholipids (1).
100
L. COCCO, et al.
The growth promoting effect of IGF-I begins at the plasma membrane with the binding to its tyrosine kinase receptor and the downstream stage of the cascade seems to be constituted by the rapid hydrolysis of nuclear polyphosphoinositides. One might argue that the mitogenic stimulation described is given by a combination of IGF-I and bombesin, the latter is capable of activating at the plasma membrane the specific PLC for PIP 2 (33). However, it is clear from the analysis of the replicative capability after mitogenic stimulation that IGF-I on its own induces DNA synthesis and bombesin acts as a synergizing agent. Moreover, only IGF-I affects both nuclear inositol lipids and PKC activity whereas bombesin is completely ineffective. The changes produced at nuclear level by IGF-I are visible only in Swiss 3T3 cells which replicate after mitogenic stimulation. Once established that the type I IGF receptor was expressed at the plasma membrane in both responsive and unresponsive cells the early changes in nuclear polyphosphoinositide levels appear to be actually causative in eliciting the onset of DNA synthesis. It is rather intriguing to establish all the steps by which the signal generated by IGF-I stimulation reaches the nucleus; however, it is well accepted that the basic operating principle is that an extracellular signal affects the activity of a protein kinase cascade and this is certainly the case for IGF-I which stimulates tyrosine kinase activity (19). Therefore, it is clear that when IGF-I binds to its receptor a phosphorylation cascade takes place and this signal reaches the nucleus. At this stage a rapid and transient hydrolysis of nuclear PIP and PIP 2 occurs and this precedes the translocation of PKC to the nucleus. The fact that this step is causative for the onset of DNA synthesis is hinted at by the evidence that only in responsive cells do nuclear polyphosphoinositide changes and PKC translocation occur. When the same mitogenic stimulus (i.e., IGF-I) is applied to unresponsive cells, which still express the IGF-I tyrosine kinase receptor, something inside the cell is imbalanced and the crosstalk between surface receptor and nuclear compartment is lacking, where the signalling system is not elicited and then DNA synthesis does not take place. The in v i t r o stimulation of nuclear PKC by means of Ca 2+ free IP 3 supports the direct link between inositol lipid hydrolysis and PKC activation. Moreover, the PLC activated hydrolysis of PIP and PIP 2 also generates DAG, whose levels at the nucleus undergo changes during cell differentiation (20) and during mitogenic stimulation (unpublished observations), so that it is conceivable that D A G can lower the threshold of nuclear PKC for Ca 2+. In addition, the PLC activated hydrolysis of PIP produces IP 2 which has been suggested to have an intranuclear function in directly stimulating DNA polymerase activity (34). Therefore, we can conclude that among IGF-I stimulated events in nuclei of Swiss 3T3 cells the changes of inositol lipids and the translocation of PKC represent a key step for the subsequent onset of DNA synthesis.
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION
101
SUMMARY T h e correlation between changes in nuclear polyphosphoinositide levels preceding P K C translocation to the nucleus and the onset of D N A synthesis has been discussed. Using two different clones of Swiss 3T3 fibroblasts belonging to the same original cell line, one of which is unresponsive to mitogenic stimulation with I G F - I on its own or in combination with bombesin, it has been observed that a rapid and transient breakdown of nuclear PIP and PIP E occurs only in responsive cells and this precedes the translocation of P K C to the nucleus, as evidenced by immunochemical analysis as well as by enzymatic activity. Therefore, it seems that a direct link exists between nuclear polyphosphoinositide metabolism, P K C translocation to the nucleus and cell division. Since I G F - I acts at the plasma m e m b r a n e through a tyrosine kinase receptor it seems that the mitogenic stimulation induced by this factor utilizes different signalling pathways at the plasma m e m b r a n e and at the nucleus. Because of the evidence that type I I G F receptor is expressed in both responsive and unresponsive cells and that the receptor machinery at the plasma m e m b r a n e is active the lack of the transient changes in nuclear inositol lipids and of P K C translocation in unresponsive cells further suggests that the cell nucleus is capable of an a u t o n o m o u s signalling system based on polyphosphoinositide metabolism. ACKNOWLEDGEMENTS The authors wish to express their appreciation to Dr. S. Capitani for his helpful advice and c o m m e n t s , to Dr. E. Faicieri for electron microscopy observations and to A. Fantazzini for his skilled technical assistance. This work was supported by Italian C . N . R . grants PF I G and PF BTBS.
REFERENCES 1. M.J. BERR1DGE and R. F. IRVINE, lnositol trisphosphate, a novel second messenger in cellular signal transduction, Nature 312,315-321 (1984). 2. C.P. DOWNES and C. H. MACPHEE, Myo-inositol metabolites and cellular signals, Eur. J. Biochem. 193, 1-18 (1990). 3. Y. NISHIZUKA, The molecular heterogeneity of protein kinase C and its implications for cellular regulation, Nature 334, 661-665 (1988). 4. M.J. BERRIDGE and R. F. IRVINE, lnositol phosphates and cell signalling, Nature 341,197-204 (1989). 5. E. ROZENGURT, Signal transduction pathways in mitogenesis. Brit. Med. Bull. 45, 515-528 (1989). 6. R. M. EVANS, The steroid and thyroid hormone receptor family, Science 240, 889-895 (1988). 7. L. COCCO, R. S. GILMOUR, A OGNIBENE, A. J. LETCHER, F. A. MANZOLI and R. F. IRVINE, Synthesis of polyphosphoinositides in nuclei of Friend cells.
102
8. 9. 10. 11. 12. 13. 14.
15.
16.
17. 18.
19. 20. 21. 22. 23. 24.
L. COCCO, et al. Evidence for polyphosphoinositide metabolism inside the nucleus which changes with cell differentiation, Biochern. J. 248, 765-770 (1987). F. A. MANZOLI, A. M. MARTELLI, S. CAPITANI, N. M. MARALDI, R. RIZZOLI, O. B A R N A B E I and L. COCCO, Nuclear polyphosphoinositides during cell growth and differentiation, Advan. Enzyme Regul. 28, 25-34 (1989). F. A. MANZOLI, S. CAPITANI and N. M. MARALDI, Minor components of chromatin and their role in the release of template restriction, pp. 463--486 in Cell Growth (C. NICOLINI, ed.), Plenum Press, New York (1982). F.A. MANZOLI, N. M. MARALDI, S. CAPITANI, L. COCCO and O. BARNABEI, Chromatin lipids and their possible role in gene expression. A study in normal and neoplastic cells, Advan. Enzyme Regul. 17, 175-184 (1979). F . A . MANZOLI, N. M. MARALDI, L. COCCO, S. CAPITANI and A. FACCHINI, Chromatin phospholipids in normal and chronic lymphocytic leukemia lymphocytes, Cancer Res. 37, 843-849 (1977). S. CAPITANI, L. COCCO, N. M. MARALDI, S. PAPA and F. A. MANZOLI, Effects of phospholipids on transcription and ribonucleoprotein processing in isolated nuclei, Advan. Enzyme Regul. 25, 425-438 (1986). L. COCCO, N. M. MARALDI, F. A. MANZOLI, R. S. GILMOUR and A. LANG, Phospholipid interactions in rat liver nuclear matrix, Biochern. Biophys. Res. Commun. 96, 890-898 (1980). S. CAPITANI, V. BERTAGNOLO, M. MAZZONI, P. SANTI, M. PREVIATI, A. ANTONUCCI and F. A. MANZOLI, Lipid phosphorylation in isolated rat liver nuclei. Synthesis of polyphosphoinositides at subnuclear level, FEBS Lett. 254, 194-198 (1989). S. CAPITANI, B. HELMS, M. MAZZONI, M. PREVIATI, V. BERTAGNOLO, K. W. A. W l R T Z and F. A. MANZOLI, Uptake and phosphorylation of phosphatidylinositol by rat liver nuclei. Role of phosphatidylinositol transfer protein, Biochim. Biophys. Acta 1044, 193-200 (1990). L. COCCO, A. M. MARTELLI, R. S. GILMOUR, A. OGNIBENE, F. A. MANZOLI and R. F. IRVINE, Rapid changes in phospholipid metabolism within nuclei of Swiss 3T3 cells induced by treatment of the cells with insulin-like growth factor I, Biochern. Biophys. Res. Commun. 154, 1266-1272 (1988). L. COCCO, A. M. MARTELLI, R. S. GILMOUR, A. OGNIBENE, F. A. MANZOLI and R. F. IRVINE, Changes in nuclear inositol phospholipids induced in intact cells by insulin-like growth factor I, Biochem. Biophys. Res. Comrnun. 159, 720-725 (1989). A . M . MARTELLI, R. S. GILMOUR, L. M. NERI, L. MANZOLI, A. N. CORPS and L. COCCO, Mitogen-stimulated events in nuclei of Swiss 3T3 cells. Evidence for a direct link between changes of inositol lipids, protein kinase C requirement and the onset of DNA synthesis, FEBS Lett. 283, 243-246 (1991). R . S . GILMOUR, C. G. PROSSER, I. R. FLEET, L. COCCO, J. C. SAUNDERS, K. D. BROWN and A. N. CORPS, From animal to molecule: aspects of the biology of insulin-like growth factor I, Brit. J. Cancer 58, 23-30 (1988). O. TRUBIANI, A. M. MARTELLI, L. MANZOLI, E. SANTAVENERE and L. COCCO, Nuclear lipids in Friend cells. Shifted profile of diacylglycerol during erythroid differentiation induced by DMSO, Cell Biol. Intl. Rep. 14, 559-566 (1990). P. NICOTERA, S. ORRENIUS, T. NILSSON and P.-O. BERGGREN, An inositol 1,4,5-trisphosphate-sensitive Ca 2÷ pool in liver nuclei, Proc. Natl. Acad. Sci. U.S.A. 87, 6858-6862 (1990). A. M. MARTELLI, R. S. GILMOUR, E. FALC1ERI, F. A. MANZOLI and L. COCCO, Mitogen-stimulated phosphorylation of nuclear proteins in Swiss 3T3 cells: evidence for a protein kinase C requirement, Exptl. Cell Res. 185, 191-202 (1989). T. P. THOMAS, H. S. T A L W A R and W. B. ANDERSON, Phorbol ester-mediated association of protein kinase C to the nuclear fraction in NIH 3T3 cells, Cancer Res. 48, 1910-1919 (1988). A . R . BUCKLEY, P. D. CROWE and D. HADDOCK-RUSSELL, Rapid activation of protein kinase C in isolated rat liver nuclei by prolactin, a known hepatic mitogen, Proc. Natl. Acad. Sci. U.S.A. 85, 8649-8653 (1988).
IGF-I, INOSITOL LIPIDS AND PKC TRANSLOCATION 25. 26. 27.
28.
29.
30. 31. 32. 33. 34.
103
A. P. FIELDS, G. TYLER, A. S. KRAFT and W. S. STRADFORD MAY, Role of nuclear protein kinase C in the mitogenic response to platelet-derived growth factor, J. Cell Sci. 96, 107-114 (1990). S. CAPITANI, P. R. GIRARD, G. J. MAZZEI, J. F. KUO, R. BEREZNEY and F. A. MANZOLI, Immunochemical characterization of protein kinase C in rat liver nuclei and subnuclear fractions, Biochem. Biophys. Res. Commun. 142, 367-375 (1987). A . M . MARTELLI, C. CARIN1, S. MARMIROLI, M. MAZZONI, P. J. BARKER, R. S. GILMOUR and S. CAPITANI, Nuclear protein kinases in rat liver: evidence for increased histone H1 phosphorylating activity during liver regeneration, Exptl. Cell Res. 195, 255--262 (1991). A . M . MARTELLI, R. S. GILMOUR, F. A. MANZOLI and L. COCCO, Calcium free inositol (1,4,5)-trisphosphate stimulates protein kinase C dependent protein phosphorylation in nuclei isolated from mitogen-treated Swiss 3T3 cells, Biochem. Biophys. Res. Commun. 173, 149-155 (1990). A. M. MARTELLI, L. M. NERI, R. S. GILMOUR, P. J. BARKER, N. S. HUSKISSON, F. A. MANZOLI and L. COCCO, Temporal changes in intracellular distribution of protein kinase C in Swiss 3T3 cells during mitogenic stimulation with insulin-like growth factor I and bombesin: translocation to the nucleus follows rapid changes in nuclear polyphosphoinositides, Biochem. Biophys. Res. Commun. 177, 480--487 (1991). O. BACHS and E. CARAFOLI, Calmodulin and calmodulin-binding proteins in liver cell nuclei, J. Biol. Chem. 262, 10786--10790 (1987). M.J. PUJOL, M. SORIANO, R. ALIGUE', E. CARAFOLI and O. BACHS, Effect of alpha-adrenergic blockers on calmodulin association with the nuclear matrix of rat liver cells during proliferative activation, J. Biol. Chem. 264, 18863-18865 (1989). C . A . ROSS, J. MELDOLESI, T. A. MILNER, T. SATOH, S. SUPAq~I'APONE and S. SNYDER, Inositol 1,4,5-trisphosphate receptor localized to endoplasmic reticulum in cerebellar Purkinije neurons, Nature 339, 468--470 (1989). J. P. HESLOP, D. M. BLAKELEY, K. D. BROWN, R. F. IRV1NE and M. J. BERRIDGE, Effects of bombesin and insulin on inositol (1,4,5) trisphosphate and inositol (1,3,4) trisphosphate formation in Swiss 3T3 cells, Cell 47, 703-709 (1986). V. SYLVIA, G. CURTIN, J. NORMAN, J. STEC and D. BUSBEE, Activation of a low specific activity form of DNA polymerase alpha by inositol-l,4-bisphosphate, Cell 54, 651~558 (1988).
