Phosphorylation, Protein Part 2 (Molecular Biology)

4. Specificity of Protein Phosphorylation

Protein kinases display a high degree of substrate specificity. First, the vast majority phosphorylate either serine/threonine or tyrosine side chains, whereas only a few kinases have a dual specificity, for both threonine and tyrosine. In addition, protein kinases are highly sensitive to the immediate environment of the phosphorylated residue (21). Thus, specific consensus sequences have been identified for many of them (Table 2). In some cases, the specificity also depends on higher-order determinants, such as interactions between the protein substrate and the protein kinase at sites located at distance from the phosphorylated residue. In living cells, the proximity between the protein kinase and its substrate plays an important role. This is achieved by specific targeting mechanisms, which enrich the kinase concentration at particular sites of the cell (22, 23). There are also a number of scaffolding proteins that associate several enzymes, including kinases and/or phosphatases and their substrates (24). In the case of protein phosphatases, the nature of the sequence surrounding the phospho-amino acid to be dephosphorylated seems generally less critical than in the case of protein kinases; and the role of higher-order determinants, as well as that of targeting processes, appears fundamental (25, 26). It should be pointed out that, with the exception of some monosubstrate systems, there is no perfect match between the substrate specificity of most protein kinases and phosphatases, implying a complex intertwining of the regulation by these enzymes.

Table 2. Examples of Consensus Phosphorylation Sites for Protein Kinase

cAMP-dependent protein kinase
Ca -calmodulin-dependent kinase II
ArgXxxXxxSer*/Thr*
Protein kinase C
Casein kinase 1
Casein kinase 2
MAP-kinases, cyclin-dependent
kinases
EGF-receptor tyrosine kinase

5. Methods to Study Protein Phosphorylation

5.1. General Methods to Study the State of Phosphorylation of Proteins

(detailed descriptions of methods can be found in Ref. 27-30). The most commonly used approach is metabolic radiolabeling of cells of interest with P-orthophosphoric acid (sometimes called "front- phosphorylation"). Pi is taken up by cells, incorporated into endogenous ATP (as well as in many other phosphorylated molecules), and used to phosphorylate proteins, lipids, and nucleic acids. Labeled phosphoproteins can be studied by electrophoresis (usually two-dimensional gel electrophoresis) or following immunoprecipitation with specific antibodies. However, proteins are often phosphorylated on several sites, which can have different physiological meanings, and other approaches are required to analyze independently the phosphorylation of each of them (see below).

Moreover, the amount of P incorporated in a given site does not reflect directly the stoichiometry of phosphorylation of this site, but rather the turnover rate of the phosphoryl group. That is, a site phosphorylated at a high stoichiometry but with a low turnover rate (eg, "constitutive phosphorylation") will appear poorly labeled, whereas a residue phosphorylated at a low stoichiometry, but with a high turnover rate, will appear intensely labeled. The levels of labeling are also dependent on the specific activity of the intracellular ATP pools, which may vary, depending on the cell population or physiological status. Therefore, other methods have been developed to avoid these pitfalls. In "back-phosphorylation" assays, proteins of interest are isolated in conditions in which their phosphorylation state is preserved. They are then phosphorylated stoichiometrically in vitro with g P-ATP and a purified kinase, active on the phosphorylation site of interest. Thus, the higher the level of phosphorylation of the site in vivo, the lesser radioactive phosphate is incorporated in vitro. The use of this method is limited, however, by the requirement for a highly active purified protein kinase and by its low sensitivity. An alternative approach that is enjoying considerable success is the use of phosphorylation-state-specific antibodies. Some of these antibodies react with a phosphorylated amino acid, almost independently of its sequence environment. For example, antibodies specific for phospho-tyrosine are extremely useful tools in the study of protein tyrosine phosphorylation. Moreover, specific antibodies can be raised against synthetic phosphopeptides encompassing the phosphorylated site of a protein of interest. Such antibodies are specific for a given protein phosphorylated at a particular site, and they provide powerful tools that can be used for immunoblotting, immunocytochemistry, and other approaches.

Phosphoproteins are often studied by one- or two-dimensional electrophoresis. In nondenaturing conditions (eg, isoelectrofocusing), the presence of a phosphorylated residue shifts the electrophoretic mobility of the protein toward a more acidic form. In the presence of SDS, phosphorylation of a protein may have no effect on its electrophoretic mobility, or it may increase its mobility and its apparent size. The slowing of the protein mobility by phosphorylation during SDS-PAGE is often attributed to a decrease in the charge density due to decreased binding of SDS to the phosphorylated protein. In contrast, it is noteworthy that phosphorylation at specific sites may increase the mobility of some rare proteins in SDS-PAGE.

5.2. Methods for Identifying the Phosphorylated Amino Acid(s)

Following prelabeling with P or in vitro phosphorylation with g P-ATP, the nature of the phosphorylated residues can be determined by acid hydrolysis of the peptide bonds and separation of phospho-serine, phospho-threonine, and phospho-tyrosine by thin-layer electrophoresis. Phospho- histidine and phospho-aspartate are extremely acid-labile and cannot be identified by this approach.

On the other hand, the resistance of phospho-tyrosine to alkaline pH is used to study this phospho- amino acid preferentially. The phosphopeptides generated by partial proteolysis of a P-labeled phosphoprotein with specific endoproteinases can be studied by two-dimensional phosphopeptide maps (usually a combination of thin-layer electrophoresis and chromatography). Such two-dimensional phospho-peptide maps are very useful for comparing the sites phosphorylated on the same protein under various circumstances, in vivo or in vitro. Phosphopeptides can also be separated by high-performance liquid chromatography (HPLC) and, if sufficient amounts of material are recovered, submitted to automatic protein sequencing, allowing the identification of the phosphorylated residue and its surrounding protein sequence. The use of powerful technologies of matrix-assisted laser desorption/ionization (MALDI) mass spectrometry now provides an interesting alternative approach to identify phosphorylation and other post-translational modifications. Site-directed mutagenesis of individual residues (replacement of a "phosphorylatable" residue by a "nonphosphorylatable" one—for example, replacement of a serine or a threonine by an alanine, or of a tyrosine by a phenylalanine ) is often used to test the phosphorylation of a precise residue. Although this approach is extremely useful, it should be kept in mind that the information obtained is indirect and that such mutations may have consequences on other properties of the protein other than preventing phosphorylation of the mutated residue.

6. Effects of Phosphorylation on the Properties of Proteins

The effects of phosphorylation on the properties of many proteins are known to be functionally important. They include the activation or inhibition of enzymes, the opening or closing of ion channels, the increase or decrease in the activity of transcription factors, the aggregation or disassembly of cytoskeletal components, and many other effects. There are only a few cases, however, in which the precise molecular mechanism by which phosphorylation of a specific residue brings about the functional changes observed is known in detail. The consequences of the phosphorylation of an amino acid residue can be only local, or they can be amplified in the three-dimensional protein structure. Phosphorylation can also alter the interactions of a protein with others. A phosphate group being bulky and highly charged (two negative charges at physiological pH), its presence alters dramatically the properties of the amino acid side chain to which it is bound. It is often attempted to mimic the consequence of phosphorylation of a residue by its mutation to a negatively charged amino acid (aspartate or glutamate). In some cases, such mutations reproduce some of the effects of phosphorylation, while they prevent the normal phosphorylation of the corresponding residues. However, the size and the charge of the carboxyl group are less than those of a phosphoryl group, accounting for the many failures of this approach.

6.1. Local Effects of Phosphorylation on the Properties of Proteins

One example of a protein in which phosphorylation modifies dramatically the activity of an enzyme by a local action on the active site is isocitrate dehydrogenase (IDH), an enzyme regulated by phosphorylation in plants and bacteria. In the unphosphorylated state, IDH is active and catalyses the oxidative decarboxylation of isocitrate to a-ketoglutarate, a critical step in the Krebs citric acid cycle. In the presence of high levels of ATP, IDH is switched off by phosphorylation, and citrate is funneled to the glyoxylate cycle, a biosynthetic pathway that allows plants and bacteria to grow on acetate. Escherichia coli IDH is phosphorylated on Ser113 by an enzyme that is also able to dephosphorylate the same residue (a tandem kinase/phosphatase; see above). Ser113 is located in the substrate binding site, near the active site of the enzyme (31). Phosphorylation of this serine prevents isocitrate binding by steric hindrance and electrostatic repulsion (Fig. 3a).

Figure 3. Effects of phosphorylation on proteins. (a) Local effects of phosphorylation on isocitrate dehydrogenase (IDH). IDH transforms isocitrate into a-ketoglutarate, with concomitant production of CO2 and reduction of NAD+ in NADH. In E. coli IDH, phosphorylation of serine-113 in the active site prevents the binding of the substrate and inactivates the enzyme. (b) Allosteric activation of glycogen phosphorylase by phosphorylation. Phosphorylase is a homodimer; phosphorylation of Ser14 in both subunits leads to rearrangement of the amino-terminal region, modification of the subunit interaction, and changes in the tertiary and quaternary structure of the enzyme, making the active site more accessible to the substrate. (c) Regulation of SH2-domain binding by tyrosine phosphorylation. When particular protein sequences are phosphorylated on tyrosine residues, they bind with high affinity to specific SH2 domains. The presence of phosphotyrosine is necessary for the binding, but the specificity of the interaction is achieved by the recognition by the SH2 domain of the residues on the carboxy-terminal side of the phosphorylated tyrosine.

Another example of a well-studied local effect of phosphorylation is provided by the structural studies of protein kinases themselves (32). For the protein kinases to be active, a peptide loop in the vicinity of the active site, termed the activation loop or T loop, must be correctly positioned. In many protein kinases, the correct positioning of the activation loop requires its phosphorylation on one or two residues. Interestingly, in some cases the phosphorylation appears constitutive (eg, protein kinase A), whereas in others (eg, mitogen-activated protein kinases, MAP kinases) the phosphorylation of the activation loop is a fundamental regulatory step, catalyzed by a specific activating protein kinase.

6.2. Effects of Phosphorylation on the Overall Structure of Proteins

Phosphorylation of a single residue can have dramatic consequences on the tertiary structure and quaternary structure of a protein, as demonstrated by the study of the X-ray crystallography structure of glycogen phosphorylase in the phosphorylated and nonphosphorylated states (33). Phosphorylase is a homodimer, and phosphorylation occurs on a serine in the amino-terminal region of each peptide chain (Ser14), located strategically at the subunit interface. Phosphorylation of this serine leads to the reorganization of the amino terminus of each subunit and the modification of its interaction with the other subunit. As a consequence, the enzyme is stabilized in a form in which the active site, located at a distance from the phosphorylated serine, becomes more accessible to the substrate (Fig. 3b). In this case, the effects of phosphorylation on the structure of the enzyme are comparable to those occurring during classical allosteric activation, except that the effects of phosphorylation are due to a covalent modification of the enzyme, and not to ligand binding.

6.3. Effects of Phosphorylation on Protein-Protein Interactions

Protein phosphorylation can have dramatic regulatory effects on protein-protein interactions, as disclosed by the study of protein tyrosine phosphorylation. Indeed, a 100-residue domain first identified in Src, after which it was termed Src-homology 2 (SH2), and then found in many other proteins, was shown to bind specific peptides phosphorylated on tyrosine (34, 35). Phosphorylation of the tyrosine residue is necessary for binding, but the specificity of the interaction is achieved by the recognition of additional residues, located on the carboxy-terminal side of the phosphorylated tyrosine, by the SH2 domain (Fig. 3c). Binding of SH2 domains to specific peptide sequences phosphorylated on tyrosine residues allows the clustering of enzymes directly or via adapter proteins, around activated growth-factors receptors. Such enzymes, including phosphatidylinositol-3-kinase, phospholipase Cg, or guanine-nucleotide exchange factors, are thus brought into the vicinity of their membrane-associated substrates. S^-mediated clustering triggers the cascades of reactions responsible for the effects of growth factors. SH2 domains can also have very different functions, such as maintaining a protein in an inactive state, as in the case of Src-family tyrosine kinases (36, 37). A carboxy-terminal phosphorylated tyrosine is involved in an intramolecular interaction with the SH2 domain, placing the enzyme in a closed conformation unable to reach its substrates. Activation results from dephosphorylation of the carboxy-terminal tyrosine, or by displacement of the SH2 domain by a competing phosphopeptide. Additional domains that interact specifically with phosphorylated proteins have been identified, including PTB domains (phosphotyrosine-binding) (38) and 14-3-3 proteins, whose interaction with their partner proteins appears to require phosphorylation of the latter on a serine residue (39).

7. The Place of Protein Phosphorylation in the Signal Transduction Networks

Many signal transduction systems in cells use several modules whose hierarchical organization allows powerful amplification and precise modulation. Protein phosphorylation is involved in most of these transduction systems, either as an output mechanism, by altering the properties of target proteins, or as a module of information processing. Indeed, phosphorylation reactions in cells often occur in cascades in which activation of a first protein kinase phosphorylates and activates a second protein kinase, which, in turn, phosphorylates and activates a third kinase, and so on. This situation may account in part for the very large number of protein kinases, and it apparently has a number of interesting properties (40, 41). First, phosphorylation cascades allow amplification of the signal: One molecule of the first activated kinase can phosphorylate several secondary kinases, which can each phosphorylate several tertiary kinases, and so on. Another property of these cascades is to acquire kinetic properties, such as positive or negative cooperativity, similar to those of allosteric enzymes, even if the kinases composing the cascade themselves follow classical Michaelis-Menten kinetics . Thus, protein kinase cascades can convert graded inputs into switch-like outputs (42). Finally, the kinase cascades provide many levels of regulation and possible cross-talks between different regulatory pathways. Thus, it is likely that the organization of protein kinases in cascades confers strong evolutionary advantages, as attested by the very high degree of conservation of many of these cascades.

8. Examples of Phosphorylation Dysfunction in Human Diseases

As mentioned above, virtually all intracellular biological processes are regulated to some extent by protein phosphorylation. Thus, it is not surprising that dysfunction of these reactions may be responsible for some pathological states, including in humans. Protein kinases and phosphatases regulate cell growth, and a number of mutations leading to the uncontrolled activation of phosphorylation pathways are oncogenic (43, 44). Several oncogenes are tyrosine kinases (eg, v-Src, v-Fyn, v-ErbB, Met) or serine/threonine kinases (eg, Raf1, Mos) that are activated by mutation, whereas it is the loss of function of protein tyrosine and phospholipiol phosphatase PTEN that is oncogenic (45). Loss of function of protein kinases is also responsible for various genetic diseases, including immune deficiency by the absence of g-globulins [mutation of a nonreceptor tyrosine kinase in Bruton's disease (46)], intestinal palsy [mutation of a receptor tyrosine kinase Ret in Hirschprung's disease (47)], or myotonic dystrophy [mutation of a serine/threonine kinase (48)]. Conversely, protein phosphatases are used as weapons by some pathogenic organisms. For example, one of the virulence genes of Yersinia bacteria, including the agent of bubonic plague, is a phosphotyrosine phosphatase (49). Some environmental toxins exert their effects by stimulating protein kinases irreversibly (eg, stimulation of protein kinase C by oncogenic phorbol esters) or inhibiting protein phosphatases (inhibition of serine/threonine phosphatases by okadaic acid or microcystins, which are responsible for food poisoning) (50). On the other hand, the immunosuppressant drugs that have allowed the recent progress in organ grafting, cyclosporin A and FK506, act by inhibiting calcineurin, a Ca /calmodulin-dependent protein phosphatase (51). Thus, the large number of protein kinases and phosphatases, and their involvement in numerous biological functions, has prompted many investigators to try to develop specific inhibitors or activators of these enzymes as potential therapeutic tools.