| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Address correspondence to (a) Dr Matthew R. Pincus,VA Medical Center, 800 Poly Place, Brooklyn, NY 11209; tel 718 630 3688; e-mail matthew.pincus2{at}med.va.gov, or to (b) Dr Sidney Pestka, RW Johnson Medical Center, 675 Hoes Lane, Piscataway, NJ; tel 732 235 4567; e-mail pestka{at}waksman.rutgers.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
(received 5 December 2002; accepted 27 March 2003)
Keywords: JAK-2-binding protein, protein arginine methyltrasferase (PRMT5), oncogenic ras-p21, insulin
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
Oncogenic, but not its wild-type counterpart, ras-p21 protein induces malignant transformation of mammalian cell lines such as NIH 3T3 cells [3]. In Xenopus laevis oocytes, microinjection of oncogenic (containing Val in place of Gly 12), but not wild-type, p21 induces oocyte maturation [4]. Insulin induces oocyte maturation and requires activation of normal cellular p21 [5]. Several agents that strongly block Val 12-p21-induced oocyte maturation have virtually no effect on insulin-induced maturation. Among these agents are specific peptides, identified from molecular modeling studies, that correspond to effector domains from both ras-p21 itself, such as the 35–47, 96–110 and 115–126 sequences [1] and from some of its target proteins such as the ras-binding domain of raf (residues 97–110) [6] and SOS guanine nucleotide exchange protein (residues 994–1004) [7].
Two targets of Val 12-p21 that appear to be unique to the oncogenic protein are jun-N-terminal kinase (JNK) and its substrate, the transcriptional activating protein, jun [8,9]. We have found that the p21 96–110 peptide blocks the interaction of Val 12-p21 with JNK, with a dose-response curve that superimposes on that for its inhibition of Val 12-p21-induced oocyte maturation [1]. The 96–110 peptide, however, has no effect on insulin-induced maturation. Since insulin-induced maturation requires activated cellular ras-p21, we concluded that direct interactions between activated wild-type ras-p21 and these target proteins do not occur.
Because both oncogenic and wild-type ras-p21 proteins require farnesylation, cleavage, and carboxyl terminal methylation, it might be expected that agents that interfere with these processes would affect the abilities of both proteins to induce mitogenesis to the same extent. In this regard, a new protein with well-defined methyl transferase activity has recently been identified in studies utilizing the two-hybrid yeast system to identify proteins that associate with Janus kinase (JAK), a signaling protein that occurs on interferon-, cytokine-, and growth factor-induced pathways [10]. This protein, called JAK-binding protein-1 (JBP-1) and more recently designated protein arginine methyl transferase-five (PRMT5), has strong homology to the yeast proteins Skb1 and Hsl7p methyl transferases, both of which utilize S-adenosyl methionine as their co-factor [11]. PRMT5 also has been found to bind to this co-factor and complements yeast strains that are defective in Hsl7p [11]. Both yeast proteins have been implicated in ras-p21 signaling in yeast, and Hsl7p, which is a functional component of the MAP kinase pathway, has been implicated in controlling the G2/M transition in the cell cycle [11]. Because of the close association of these yeast proteins and, presumably PRMT5, with the ras-p21 signaling pathway and their association with control of the cell cycle, we tested whether PRMT5 has an effect on oncogenic and normal ras-p21-induced oocyte maturation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
Vectors and mRNA. Vectors encoding JBP1 (PRMT5) and 2 mutants, the R368A substitution mutant (PRMT5-M1) and the GAGRG (365–369) deletion mutant (PRMT5-M2), were prepared as described previously [10,11,13]. Briefly, for native PRMT5, the full-length PRMT5 cDNA was cloned as a SalI fragment into the compatible XhoI site of bluescript SK- to yield p41-6 bluescript. The mutants of PRMT5 were constructed using PCR with the appropriate oligonucleotides: 41Del-1 and MUT1 for the R368A mutant and 41 Del-1 and the deletion sequence for the GAGRG deletion mutant and pEF2-JBP1 as template. Each construct was linearized and transcribed using an in vitro transcription kit (Riboprobe Transcription System, Promega, Madison, WI). All mRNAs were isolated as described previously [13,14], washed several times with ethanol, and stored at -80°C until used. The ß-gal empty vector was prepared as described previously [14].
Oocyte microinjection. Oocytes were obtained from Xenopus laevis frogs (Connecticut Valley Biological, Southhampton, MA) as described previously [6,7]. All microinjection experiments were performed at least 6 times independently on 30 oocytes, prepared from collagenase-digested ovarian follicles that were then incubated at 19°C for 12–24 hr. Microinjected oocytes were incubated in Barth’s medium, or Barth’s medium containing insulin (10 µg/ml), for 36 or 48 hr at 19°C. Oocyte maturation was determined by observing germinal vesicle breakdown (GVBD) (white patches on the brown animal pole) as previously described [4,5].
[Val 12]Ha-ras-p21 was injected alone or co-injected with the appropriate mRNA encoding the PRMT5 protein. Negative controls included ß-Gal mRNA, albumin, and the X-13 peptide. Each mRNA encoding a specific PRMT5 protein was injected by itself, together with Val 12-p21 (100 µg/ml), or into oocytes subsequently incubated with insulin. In all experiments involving microinjection of potential inhibitory agents with insulin as the agent inducing maturation, insulin was added to the incubation medium 1 hr after microinjection.
Injection of methyl transferase inhibitors. Incubation of cells with adenosine and D,L-homocysteine in the presence of the S-adenosyl-homocysteine (SAH)-hydrolase inhibitor, N-methyl-2-deoxyadenosine, results in the cellular accumulation of SAH, a potent inhibitor of protein arginine-methyltransferases [15, and Zhu W, Mustelin T, and David M. Arginine methylation of STAT1 regulates its dephosphorylation by TcPTP, submitted for publication (personal communication)]. Therefore, to test whether SAH inhibits oocyte maturation induced either by microinjected Val 12-p21 or by insulin, oocytes were injected either with Val 12-p21 (100 µg/ml) together with a methyl transferase incubation mixture consisting of adenosine (300 µM), homocysteine (300 µM) and N-methyl-2-deoxyadenosine (300 µM), or with the inhibition mixture alone and then incubated with insulin.
Statistical methods. In all sets of experiments, the maturation after either 36 or 48 hr was expressed as mean ± SD. Differences between the maturation in different groups of experiments were tested for statistical significance by two-tailed t-test. In the Results section, all differences between average maturations in 2 groups that are significant have p values <0.001; those that are not significant have p values 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
shows typical time courses for induction of oocyte maturation by ras-p21 and insulin. Each of these agents is seen to induce about the same level of maturation over a 36-hr time course, suggesting that both agents are present at functionally equivalent concentrations. Co-injection of Val 12-p21 with bovine serum albumin, the unrelated control peptide X13, or empty ß-gal vector resulted in no inhibition (not shown). In contrast (Fig. 1), PRMT5 protein strongly inhibits oncogenic p21-induced oocyte maturation (51.7% ± 2.7% with p21 alone vs 6.9 ± 2.5% with p21 + PRMT5, p <0.001). PRMT5 has much less effect on insulin-induced maturation (52.8 ± 4.5% with insulin alone vs 48.8 ± 0.9, p >0.05 with insulin + PRMT5). The mutant PRMT5-M1, containing an R368A substitution that reduces but does not abolish methyl transferase activity [10], causes less inhibition of oncogenic p21-induced maturation but has increased inhibitory activity on insulin-induced maturation. Interestingly, PRMT5 and, to a slightly lesser extent, PRMT5-M1, actually induce oocyte maturation (Fig. 1 and conditions 7 and 8 in Fig. 2).
|
|
. Inhibition of maturation induced by oncogenic p21 by the two PRMT5 proteins (conditions 2 and 3) appears to correlate with their methyl transferase activities. PRMT5 has full methyl transferase activity while the R368A-substituted protein has diminished but significant residual activity [10]. As shown in conditions 2 and 3 in Fig. 2, PRMT5 more strongly blocks oncogenic p21-induced maturation (condition 2) than does its less active counterpart (condition 3). These results suggest that methylation of oncogenic ras-p21 or proteins in the ras-p21 pathway causes its inactivation. The fact that native PRMT5 has only minimal effect on insulin-induced maturation (condition 5) suggests that inhibition by PRMT5-M1 of insulin-induced oocyte maturation is at least partially independent of its methyl transferase activity. This is in contradistinction to the results with oncogenic p21 protein.
Absence of inhibition of insulin-induced maturation occurred irrespective of whether the PRMT5 proteins were injected into oocytes just prior to incubation with insulin or from 1–3 hr prior to the addition of insulin to the incubation medium. This indicates that lack of inhibition of insulin-induced maturation by both proteins is not due to insufficient time for them to diffuse to their sites of inhibition. Since absence of inhibition is independent of the time of injection prior to addition of insulin and since both proteins inhibit Val 12-p21 over a 36-hr time course, it is unlikely that absence of inhibition of insulin is due to degradation of PRMT5.
Effects of a totally inactive PRMT5 mutant. To investigate further the correlation between inhibition of oocyte maturation and methyl transferase activity, we prepared a mutant PRMT5 protein, PRMT5-M2 [13], in which the active site sequence, GAGRG (residues 365–369), was deleted, resulting in low methyl transferase activity. We tested it for the ability to inhibit oocyte maturation induced either by oncogenic p21 or by insulin.
Fig. 3 shows the effect of this mutant protein on oncogenic p21- and insulin-induced maturation. Condition 2 in this figure shows that PRMT5-M2 has much less effect on oncogenic p21-induced maturation than does native PRMT5 protein (condition 2, Fig. 2). Furthermore, it has less inhibitory effect on oncogenic p21-induced maturation (40% maturation) than does the "leaky" PRMT5-M1 protein (20% maturation, Fig. 2). This supports the correlation between methyl transferase activity and extent of inhibition of ras-p21-induced oocyte maturation.
|
). These proteins differ in that PRMT5-M1 has an Ala-for-Arg 368 substitution, while in PRMT5-M2, the Gly-Ala-Gly-Arg-Gly (365–369) sequence is deleted. This sequence represents the Ado Met binding site. This result suggests that residues 365–369 may be important in the inhibitory activity of PRMT5-M1 on insulin-induced maturation. We therefore either co-injected the tripeptide Gly-Arg-Gly with Val 12-ras-p21 into oocytes or injected this peptide into oocytes incubated with insulin.
As shown in Fig. 3 (condition 7), Gly-Arg-Gly peptide strongly blocks insulin-induced maturation but has no effect on Val 12-ras-p21-induced maturation (condition 3). This supports the importance of the local tripeptide sequence on insulin-, but not on oncogenic ras-p21-induced maturation.
Effects of methyl transferase inhibitors on oocyte maturation. Since carboxymethylation of ras-p21 is a necessary step in its post-translational processing [2], we explored the effect of methyl transferase inhibitors on the activities of oncogenic ras-p21 protein and insulin in induction of oocyte maturation. To this end, we injected a mixture of methyl transferase inhibitors into oocytes either together with Val 12-p21 or incubated with insulin. This mixture consists of adenosine and D,L-homocysteine together with the S-adenosyl-homocysteine (SAH)-hydrolase inhibitor, N-methyl-2-deoxyadenosine. This results in the cellular accumulation of S-adenosylhomocysteine, a potent inhibitor of protein arginine-methyltransferases [15, and Zhu W, Mustelin T, and David M. Arginine methylation of STAT1 regulates its dephosphorylation by TcPTP, submitted for publication (personal communication)].
As shown in Fig. 3 (condition 4), the methyl transferase inhibitors block oncogenic ras-p21-induced maturation but, as shown in condition 8, have only minimal effect on insulin-induced maturation. These results support the critical importance of methyl transferase activity to the activity of oncogenic p21, but not to that of insulin. These results also imply that oncogenic p21 requires intracellular methyltransferase activity but, paradoxically, is blocked by a methyltransferase, PRMT5. This transferase, therefore, is unlikely to be involved in the post-translational carboxymethylation process required for activation of ras-p21. The site of action of PRMT5 is probably farther downstream of the oncogenic ras-p21 pathway; the target of PRMT5 appears to be unique to the oncogenic protein since it has no effect on insulin-induced maturation.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
Injection of PRMT5 mRNA into oocytes strongly blocks oncogenic ras-p21-induced oocyte maturation in a manner that suggests that the inhibition is methyltransferase-dependent. The Inhibition by wild-type PRMT5 is almost complete (condtion 2, Fig. 2) while minimal inhibition occurs with methyltransferase-deficient PRMT5-M2 that has Ado Met binding site residues 365–369 deleted (condition 2, Fig. 3). On the other hand, partially inactivated R368A PRMT5-M1 exhibits partial inhibition of oncogenic p21-induced maturation (condition 3, Fig. 2).
Paradoxical effects of PRMT5.
As shown in Figs. 1 and 2, PRMT5 and, to a lesser extent PRMT5-M1, induce oocyte maturation by themselves even though each inhibits oncogenic p21. Similar effects have been observed with peptides, such as the p21 96–110 sequence, that inhibit oncogenic ras-p21-induced oocyte maturation. These results suggest that these ras-p21 inhibitors act as partial agonists. Thus, the inhibitory 96–110 p21 peptide may induce oocyte maturation, which occurs more slowly and to a lower overall extent than that induced by Val 12-p21, because of its ability to mimic in structure one of several ras effector domains involved in mitogenic signal transduction. It is this mode that may nonetheless block maturation induced by the whole Val 12-p21 protein.
The mode of inhibition of oncogenic p21 by PRMT5 appears more complex, because it does not contain obvious ras effector domains that could compete for binding to ras-p21 targets. Thus PRMT5 may inactivate, either directly or indirectly, ras-p21 and/or vital ras-p21 targets.
The actual substrates for PRMT5 methyltransferase in oocytes are unknown. It is likely that myelin basic protein and several small ribonuclear proteins of the RNA splicing components (SMD1, SMD3, SMB1) are physiological substrates, since these proteins are symmetrically dimethylated. PRMT5 is the only known protein arginine methyltransferase to perform symmetrical dimethylation of arginine; it carries out symmetrical dimethylation of these proteins in vitro [13]. We have obtained data that suggest that PRMT5 can directly methylate Val 12-p21 in vitro to a small extent (Lee J-H, Cook JR, Kim Y, Izatova L, Pestka S, Pincus MR and Brandt-Rauf PW, manuscript in preparation), possibly implicating ras-p21 itself as a target. Whether or not this would be an inactivating event for oncogenic p21 is unknown.
If PRMT5 inactivates ras-p21 and/or one or more ras-p21 targets, it remains to be explained why PRMT5 does not continue to induce maturation (Fig. 1). One possible explanation is that, if PRMT5 inactivates ras-p21 and/or ras-induced targets and if these targets inactivated by PRMT5 do not lie on the PRMT5 signal transduction pathway, they may block PRMT5 as competitive inhibitors.
Paradoxically, oncogenic p21-induced maturation, which is blocked by PRMT5 methyl transferase, is also blocked by methyl transferase inhibitors (condition 4, Fig. 3). This suggests that there are other intracellular methyl transferases that are critical in maintaining the functional integrity of the ras-p21 pathway. Since PRMT5 occurs intracellularly, it may be activated to block oncogenic p21-induced signaling as a regulatory event or may itself be blocked through other regulatory pathways induced by the oncogenic p21 protein.
In contrast, insulin-induced maturation, which requires activation of cellular ras-p21 [5], is not affected by PRMT5 in a manner that correlates with its methyltransferase activity. It is weakly blocked by PRMT5 (condition 5, Fig. 2), is somewhat more strongly inhibited by partially activated PRMT5-M1 (condition 6, Fig. 2) and is only weakly inhibited by totally inactivated methyltransferase PRMT5-M2 mutant protein with residues 365–367 deleted (condition 6, Fig. 3). These findings suggest that inhibition of insulin-induced oocyte maturation by PRMT5 does not depend upon its methyltransferase activity. In fact, insulin-induced maturation appears to have no dependence on protein arginine-methyltransferase activity at all, since the arginine-methyltransferase inhibitor mixture had no effect on insulin-induced maturation (condition 8, Fig. 3). Since all ras-p21 molecules apparently require carboxymethylation as a post-translational event [2], it can be surmised that this process occurs by non-arginine methyltransferases.
These results support our prior conclusion that activated wild-type and oncogenic forms of ras-p21 induce different mitogenic signal transduction pathways. PRMT5 clearly shows strong specificity for blocking oncogenic rather than insulin-activated wild-type ras-p21-induced oocyte maturation.
Implications of unique peptide inhibition of insulin-induced maturation for signal transduction pathways. Surprisingly, the tripeptide sequence, Gly-Arg-Gly, from the substrate site for PRMT5 and some other protein arginine methyltransferases, blocks insulin- rather than oncogenic p21-induced maturation. In prior studies, we found that 3 peptides from ras-p21, viz 35–47, 96–110 and 115–126, selectively block oncogenic ras-p21-induced oocyte maturation [1]. Based on molecular modeling studies that identified putative effector domains of target proteins of p21, i.e., raf and SOS proteins [6,7], we have synthesized peptides corresponding to these domains and have found that they either block both oncogenic ras-p21- and insulin-induced maturation or uniquely oncogenic p21-induced maturation. Recently, in parallel studies, we have found that 2 peptides from GTPase activating protein (GAP) more selectively block insulin-induced maturation, but also inhibit oncogenic p21 appreciably though to a lesser extent [16]. Only one agent, phosphatase 2A, showed selectivity for blocking insulin-induced maturation [17]. Therefore, our finding of selective inhibition of insulin by the Gly-Arg-Gly peptide appears unique; this is the first peptide that shows a strong preference for inhibiting insulin-induced maturation.
This peptide may be useful in elucidating insulin-activated pathways that function independently of the oncogenic protein. The sites of inhibition of insulin-induced maturation may either be targets of activated cellular wild-type ras-p21 or other insulin-activated signal transduction targets on synergistic but ras-independent pathways.
Absence of inhibition (or enhancement) of insulin-induced maturation by PRMT5 and the weak inhibition by PRMT5-M1 may signify that insulin does not require PRMT5 for its activity. Further, while the Gly-Arg-Gly peptide corresponds to the sequence of the methyl transferase substrate site for PRMT5, it is a small peptide, and its targets may not necessarily be related to PRMT5 function.
Another possible explanation for inhibition of insulin-induced maturation by Gly-Arg-Gly, partial inhibition by PRMT5-M1, but no inhibition by native PRMT5, is that insulin may require PRMT5 as a signal transduction target. Since methyl transferase inhibitors do not block insulin-induced maturation, this dependence would be unrelated to PRMT5 methyl transferase activity. Under this scenario, partial inhibition by PRMT5-M1 could be due to a weak dominant negative competitive effect. The Gly-Arg-Gly peptide could likewise compete for binding to PRMT5 targets (methylation substrates), resulting in signal transduction blockade. Since the PRMT5 365–369 deletion mutant lacks the Gly-Ala-Gly-Arg-Gly of the S-Ado Met binding domain, it would have a minimal effect on insulin-induced maturation since it would presumably not interact competitively with PRMT5 targets.
|
Acknowledgements. |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionAcknowledgements.References |
|---|
This article has been cited by other articles:
|
|
 |
|
  J.-H. Lee, J. R. Cook, Z.-H. Yang, O. Mirochnitchenko, S. I. Gunderson, A. M. Felix, N. Herth, R. Hoffmann, and S. Pestka PRMT7, a New Protein Arginine Methyltransferase That Synthesizes Symmetric Dimethylarginine J. Biol. Chem., February 4, 2005; 280(5): 3656 - 3664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Singh, A. P. Sowjanya, and G. Ramakrishna The wild-type Ras: road ahead FASEB J, February 1, 2005; 19(2): 161 - 169. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
