A New Role for Protein Methylation: Switching Partners at the Phosphatase Ball

See allHide authors and affiliations

Science's STKE  24 Apr 2001:
Vol. 2001, Issue 79, pp. pe1
DOI: 10.1126/stke.2001.79.pe1


Reversible protein methylation may be another posttranslational modification that serves an important role in modulating signal transduction pathways. Not only does protein phosphatase 2A (PP2A), a key regulator of many signal transduction cascades, undergo this modification, but the methylation process itself may be regulated by various cellular stimuli or states. Mumby describes how PP2A is regulated and how methylation may alter its subcellular targeting and substrate specificity by influencing its interaction with regulatory subunits.

Reversible covalent modification of proteins is a cornerstone in cellular signal transduction. Known regulatory modifications include phosphorylation, acetylation, lipid modifications, glycosylation, and methylation. Covalent modifications regulate several aspects of signaling molecule function, but are especially important in regulating protein-protein interactions and formation of signaling complexes. A classic example is the recruitment of SH2 domain-containing signaling molecules to tyrosine-phosphorylated proteins. Although phosphorylation and dephosphorylation are the most prevalent protein modifications, several of the less abundant modifications, including methylation, also play critical roles in cell signaling. Methylated signaling molecules include low molecular weight guanosine triphosphatases (GTPases) Ras, Rac, Rho, and Cdc42, the gamma subunit of heterotrimeric GTP-binding proteins (G proteins), and the catalytic subunit of protein phosphatase 2A (PP2A). Although methylation plays a key role in sensory adaptation in bacteria, the role of reversible protein methylation in eukaryotic cells has been obscure. This gap in our knowledge is beginning to narrow. Several recent papers demonstrate that methylation acts as a molecular switch that controls the assembly of PP2A holoenzymes (1-4). These studies suggest that, like phosphorylation, methylation regulates protein-protein interactions and the recruitment of regulatory proteins into PP2A complexes. Because of the widespread involvement of PP2A in phosphorylation-dependent signal transduction, methylation has important ramifications for cell signaling.

Once thought of as a single, broad-specificity phosphatase, PP2A is actually many different enzymes composed of a common core enzyme (core dimer) that associates with regulatory subunits and other interacting proteins. These interacting proteins target PP2A to specific substrates and intracellular locations, thereby controlling its functions (5, 6). The PP2A core dimer is a high-affinity complex (AC) between a scaffold protein (the A subunit) and the catalytic subunit (C) (Fig. 1) (7). Association of regulatory subunits and interacting proteins with the core dimer generates a variety of heterotrimeric holoenzymes that account for the ability of PP2A to participate in many aspects of signaling. The PP2A catalytic subunit contains a phosphatase domain that is conserved with other members of the protein serine-threonine phosphatase gene family (8, 9). Regions outside the conserved phosphatase domain have little similarity to other members of the family [for example, protein phosphatase-1 and calcineurin (also known as PP2B)], and are responsible for interactions with PP2A-specific regulatory proteins. The PP2A-unique COOH-terminal region is important for interaction of the core dimer with the regulatory subunits. The six COOH-terminal residues (TPDYFL) are conserved in all PP2A catalytic subunits, and the last three amino acids (YFL) are conserved in protein phosphatases 4 and 6.

Fig. 1.

Reversible methylation regulates the assembly and distribution of PP2A holoenzymes. Differential association with various regulatory subunits and targeting molecules regulates the activity, specificity, and localization of PP2A. The PP2A catalytic subunit (C) forms complexes with a PP2A-specific scaffold protein, the A subunit (A). The resulting pool of AC dimers is the central intermediate in the equilibrium controlling the distribution of PP2A isoforms. In some cells, an alternative pathway for the catalytic subunit is the formation of complexes with the α4 protein, which competes with the A subunit. The core dimer is methylated (Me) by a specific protein phosphatase methyltransferase (PPMT). Methylation enhances the affinity of the AC dimer for PP2A regulatory subunits (R). The AC dimer is demethylated by a specific protein phosphatase methylesterase (PPME). The AC dimer can also interact with a variety of proteins (X) that bind independently of catalytic subunit methylation. Changes in the activities of PPMT or PPME would alter the equilibrium between different PP2A holoenzymes by changing the relative affinity of the AC dimer for methylation-dependent regulatory subunits. Because various holoenzymes have different localizations and substrate specificities, altered methylation would change PP2A activity. Increased methylation would increase activity toward substrates targeted by the methylation-dependent regulatory subunits, whereas decreased methylation would enhance dephosphorylation of substrates targeted by methylation-independent interacting proteins.

The catalytic subunit of PP2A is reversibly carboxymethylated on the COOH-terminal leucine residue (10, 11). Methylation is catalyzed by a unique protein phosphatase methyltransferase (PPMT) that is specific for the catalytic subunits of PP2A, PP4, and possibly PP6 (10, 12). PPMT contains a conserved S-adenosyl methionine-binding motif, but is distinct from other protein methyltransferases. Humans and yeast have two related PPMT genes, only one product of which (PPM1) methylates PP2A (2, 13). Although its substrates are unknown, the presence of a second methyltransferase suggests that other proteins are carboxymethylated in a manner analogous to that of PP2A. The core dimer, not the free catalytic subunit, is the preferred substrate for PPMT (1). The methyl group is removed by a specific methylesterase (14). This protein phosphatase methylesterase (PPME) was first identified by its ability to form a complex with an inactive mutant of the catalytic subunit (15). Neither regulatory subunits nor other interacting proteins are in the complexes with PPME and the core dimer. Therefore, like PPMT, PPME appears to selectively interact with, and demethylate, the core dimer and not the free catalytic subunit. The regulatory subunits compete with PPME and protect the holoenzymes from demethylation (1). Methylation does not directly alter the activity of the PP2A catalytic subunit. Methylation has no direct effect on the activity of the core dimer, but rather alters enzyme activity by changing the composition of oligomeric PP2A complexes (1, 13).

The first hints that methylation might regulate the assembly of PP2A heterotrimers came from studies with methylation-deficient mutants of the catalytic subunit in mammalian cells. Deletion of the COOH-terminus and, consequently, the site of methylation prevents the binding of the regulatory subunits to the core dimer (16). Expression of PP2A catalytic subunits containing mutations of the COOH-terminal leucine to nonmethylatable amino acids results in recovery of the mutant proteins exclusively in core dimers and not as PP2A holoenzymes (17, 18). The data from mutants are also supported by data comparing methylation of wild-type PP2A isoforms. Most of the heterotrimeric PP2A isolated from either transfected COS cells (18) or bovine brain (1) contains methylated catalytic subunit, while the core dimer isolated from the same sources is not methylated.

Tolstykh et al. (1) show that methylation enhances the affinity of the core dimer for regulatory subunits. Incubation of purified core dimers with purified PPMT and its substrate, S-adenosyl methionine, results in methylation of the catalytic subunit. When methylated core dimers are mixed with purified B/R2 (one of the three major families of PP2A regulatory subunit), substantially greater amounts of heterotrimer are recovered than with nonmethylated core dimers. Although methylation enhances the affinity of the core dimer for regulatory subunits, the effect is not absolute. AC-B/R2 trimers can still form in the absence of methylation, but the nonmethylated trimers have reduced stability. Methylation also increases the affinity of the core dimer for a second family of PP2A regulatory subunit (the B(/R5 family) (1).

Methylation of the PP2A catalytic subunit selectively enhances association of the core dimer with a subset of PP2A-interacting proteins. The B/R2 and B′/R5 regulatory subunits are two examples of a growing set of proteins that associate with the PP2A core dimer (5, 6). Although the effect of methylation has not been tested in a systematic way, the binding of some novel interacting proteins is not affected by methylation. In contrast to the effect on the interactions with regulatory subunits, lack of methylation has no effect on the association of polyomavirus middle tumor antigen, striatin, and S/G2 nuclear autoantigen with the core dimer (4). The existence of methylation-dependent and methylation-independent interacting proteins has important consequences. Increased methylation of the core dimer would promote association with methylation-dependent proteins and decrease the formation of complexes with methylation-independent interacting proteins. As discussed below, shifts in the equilibrium between core dimers and various holoenzymes would alter targeting of PP2A and subsequent protein dephosphorylation.

The idea that reversible methylation regulates the distribution of the PP2A between core dimers and different heterotrimeric holoenzymes implies that there must be a pool of free core dimers. PPMT acts only on free core dimers, and incorporation of methylated dimers into holoenzymes blocks hydrolysis by PPME. A change in the activity of PPMT or PPME would only alter the distribution of free core dimers between methylated and nonmethylated forms. An important point in further defining the role of methylation is determining the relative amounts of core dimer and holoenzymes present in cells. This has been a thorny issue because resolution of different PP2A oligomers generally requires chromatographic separation, which can cause artifactual dissociation of holoenzymes. Although core dimers are generally present in all PP2A preparations, PP2A in muscle extracts can be isolated under conditions where very little core dimer is detected (19, 20). Thus, one prevalent notion is that most of the PP2A in cells or tissues is present as heterotrimeric holoenzymes. If this were the case, then the pool of core dimer would be derived from dissociation of heterotrimers and would be small, and methylation-induced changes in holoenzyme distribution would require dissociation of holoenzymes. However, there are data indicating that a substantial pool of core dimer exists. Two-thirds of the PP2A in erythrocyte cytosol is present as the core dimer (21). Analysis with monoclonal antibodies specific for different complexes shows that, under conditions that minimize proteolysis and dissociation, one-third of the total PP2A in cell lysates is present as core dimers (22). Thus, there is compelling evidence that a substantial pool of free core dimer is present in most cells. This pool of core dimer may be present in excess of the interacting proteins, such that the core dimer is never limiting. This argues that changes in PP2A methylation and demethylation can occur rapidly and cause a substantial shift of a preexisting pool of core dimer between methylated and nonmethylated forms.

Genetic analysis in yeast supports an important role for methylation in PP2A heterotrimer formation. Inactivation of the PP2A methyltransferase gene PPM1 in Saccharomyces cerevisiae, or expression of a nonmethylatable catalytic subunit mutant, decreases the recovery of PP2A heterotrimers containing the yeast B/R2 subunit encoded by the CDC55 gene (2). Inactivation of PP2A methyltransferase also decreases the amount of A subunit in complexes with the catalytic subunit. Loss of ppm1 methyltransferase also reduces the amount of the S. cerevisiae B′/R5 subunit (RTS1) in PP2A complexes (3). Thus, interaction of both families of regulatory subunit (B/R2 and B′/R5) with the core dimer is enhanced by methylation in vivo. In contrast, other data provide evidence that interaction of the yeast B′/R5 subunit with the core dimer is not altered by mutation of the COOH-terminal amino acid to a nonmethylatable amino acid (23). Loss of PP2A methyltransferase may have additional effects, beyond decreased catalytic subunit methylation, that lower the stability of PP2A oligomers.

Blockade of PP2A methylation in yeast causes a set of phenotypes that are consistent with decreased formation of PP2A holoenzymes. In S. cerevisiae, formation of PP2A holoenzymes is not necessary for growth on rich medium. Consistent with the dispensability of holoenzymes, decreased PP2A methylation has no effect on growth in rich medium (2, 3, 23). On the other hand, formation of holoenzymes is necessary for growth under various types of stress, and loss of individual regulatory subunits leads to a number of defects. Inactivation of the gene for the yeast B/R2 subunit causes cold sensitivity, defects in cytokinesis, defects in the mitotic spindle checkpoint, and decreased sensitivity to the macrolide antibiotic rapamycin (24-27). Inactivation of the yeast gene for the B′/R5 subunit causes a distinct set of phenotypes, including temperature-sensitivity, deficient growth on nonfermentable carbon sources, and defects in the G2 to M transition (28). A careful analysis of yeast strains deleted for the PP2A methyltransferase shows that loss of methylation causes defects similar to some, but not all, of the defects caused by inactivation of the B/R2 and B′/R5 regulatory subunits (2). In contrast to B/R2 and B′/R5 deletion strains, ppm1Δ strains do not have a cold-sensitive defect in cytokinesis and do not display defects in growth on nonfermentable carbon sources. On the other hand, ppm1Δ strains do show a resistance to rapamycin similar to that of B/R2 mutants and a temperature-sensitive phenotype similar to that of B′/R5 mutants. Inactivation of PPM1 also causes defects in the spindle checkpoint similar to those of B/R2 regulatory subunit mutants. The mitotic spindle checkpoint defect is also seen in other PPM1-deficient strains (3) and strains expressing a nonmethylatable mutant of the PP2A catalytic subunit (23). Consistent with an important role for methylation, overexpression of the PP2A methylesterase gene PPE1 causes phenotypes similar to those seen with loss of the PPM1 methyltransferase (2). These observations are all consistent with a decreased ability of nonmethylated catalytic subunits to form heterotrimeric complexes with B/R2 and B′/R5 subunits. The less severe phenotypes of the methylation-deficient strains, compared to those of strains in which the regulatory subunit is deleted, are consistent with the observation that methylation enhances the affinity for regulatory subunits, but holoenzymes can still form in the absence of methylation (1). The ability of heterotrimers to form complexes with nonmethylated core dimers, albeit with reduced efficiency, is also supported by the observation that overexpression of the yeast B/R2 subunit reverses the rapamycin-resistant phenotype of ppm1 deletion strains (2).

Demonstration of a physiological role for methylation necessitates a new paradigm for understanding how protein-protein interactions are controlled in PP2A (Fig. 1). The scaffold (A) and catalytic (C) subunits assemble into the core dimer immediately after synthesis. An alternative pathway is the assembly of the catalytic subunit with the α4 protein in cells where this alternative to the A subunit is expressed. The core dimer (AC) then forms complexes with the repertoire of regulatory subunits and interacting proteins expressed in a particular cell type. The relative levels of individual holoenzymes are regulated primarily by equilibrium thermodynamics. Interacting proteins with the highest affinities and those present at the highest concentrations will dominate the population of PP2A present in a given cell type. Methylation affects this equilibrium by increasing the affinity of the core dimer for specific classes of interacting proteins (currently known to be two families of regulatory subunits). Methylation-induced changes in holoenzyme configuration would alter dephosphorylation of subsets of proteins via altered substrate targeting. Important questions remain as to whether the activities of PPMT or PPME are regulated, and whether the level of PP2A methylation changes in response to cell stimulation or adaptation to new conditions. There are a few hints that PP2A methylation is dynamic. PP2A methylation changes during the cell cycle in fibroblasts (29) and adenosine 3′,5′-monophosphate (cAMP) increases PP2A methylation in frog oocyte extracts (30). These observations suggest that the PP2A methyltransferase or methylesterase is subject to regulation. Reversible methylation has emerged as a novel mechanism for controlling the protein-protein interactions that regulate PP2A function. This regulatory paradigm for methylation may be applicable to other signaling proteins where functions for methylation have not yet been established (31).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
View Abstract

Navigate This Article