Identification of novel ERK2 substrates through use of an engineered kinase and ATP analogs

The mitogen-activated protein kinases are key regulators of cellular organization and function. To understand the mechanisms(s) by which these ubiquitous kinases affect specific cellular changes, it is necessary to identify their diverse and numerous substrates in different cell contexts and compartments. As a first step in achieving this goal, we engineered a mutant ERK2 in which a bulky amino acid residue in the ATP binding site (glutamine 103) is changed to glycine, allowing this mutant to utilize an analog of ATP (cyclopentyl ATP) that cannot be used by wild-type ERK2 or other cellular kinases. The mutation did not inhibit ERK2 kinase activity or substrate specificity in vitro or in vivo. This method allowed us to detect only ERK2-specific phosphorylations within a mixture of proteins. Using this ERK2 mutant/analog pair to phosphorylate ERK2-associated proteins in COS-1 cells, we identified the ubiquitin ligase EDD (E3 identified by differential display) and the nucleoporin Tpr (translocated promoter region) as two novel substrates of ERK2, in addition to the known ERK2 substrate Rsk1. To further validate the method, we present data that confirm that ERK2 phosphorylates EDD in vitro and in vivo. These results not only identify two novel ERK2 substrates but also provide a framework for the future identification of numerous cellular targets of this important signaling cascade.     

Conserved docking site is essential for activation of mammalian MAP kinase kinases by specific MAP kinase kinase kinases

Mammalian mitogen-activated protein kinase (MAPK) cascades control various cellular events, ranging from cell growth to apoptosis, in response to external stimuli. A conserved docking site, termed DVD, is found in the mammalian MAP kinase kinases (MAPKKs) belonging to the three major subfamilies, namely MEK1, MKK4/7, and MKK3/6. The DVD sites bind to their specific upstream MAP kinase kinase kinases (MAPKKKs), including MTK1 (MEKK4), ASK1, TAK1, TAO2, MEKK1, and Raf-1. DVD site is a stretch of about 20 amino acids immediately on the C-terminal side of the MAPKK catalytic domain. Mutations in the DVD site strongly inhibited MAPKKs from binding to, and being activated by, their specific MAPKKKs, both in vitro and in vivo. DVD site mutants could not be activated by various external stimuli in vivo. Synthetic DVD oligopeptides inhibited specific MAPKK activation, both in vitro and in vivo, demonstrating the critical importance of the DVD docking in MAPK signaling.     

Strategies for the identification of kinase substrates using analog-sensitive kinases

Phosphorylation of proteins is a prevalent post-translational modification, which affects intracellular signaling in many ways. About 2% of all eukaryotic genes code for protein kinases catalyzing phosphorylation events. Despite technological advances that have made it possible to identify thousands of phosphorylation sites simultaneously, identification of the substrates of a given kinase remains an exceptionally challenging task. Here, we summarize approaches for substrate identification that make use of genetically engineered ‘analog-sensitive’ kinases.     

Phosphorothioate analogs of ATP as the substrates of dynein and ciliary or flagellar movement

The phosphorothioate analog of ATP has a sulfur atom replacing a non-bridging oxygen atom of the triphosphate moiety of ATP. Due to the tetrahedral nature of the phosphorus atom, stereoisomers are known to exist, designated as the Sp and Rp isomers. We have reported [Shimizu & Furusawa (1986) Biochemistry 25, 5787] on the hydrolytic activity of the 22S dynein from Tetrahymena cilia towards the phosphorothioate analogs of ATP. In this paper, we extend our study and report on the microtubule-dynein dissociation by these analogs and on their ability to support sea urchin flagellar dynein enzymatic activity as well as ciliary or flagellar motility. It has been shown that the microtubule--22S-dynein complex is dissociated by the binding of ATP to the dynein enzymatic sites [Porter & Johnson (1983) J. Biol. Chem. 258, 6575]. We studied the dissociation by adenosine 5'-[alpha-thio]triphosphate (ATP[alpha S]), Sp or Rp, by light-scattering stopped-flow methods. The dissociation by (Sp)ATP[alpha S] was rapid and the rate of the light-scattering change by (Sp)ATP[alpha S] was a hyperbolic function of the nucleotide concentration, indicating that dissociation was a two-step process. On the other hand, (Rp)ATP[alpha S] up to 2 mM induced only slow and partial dissociation of the complex, while, in the presence of vanadate, it induced complete dissociation with a slightly higher rate (0.5 s-1). The adenosine 5'-[beta-thio]triphosphate (ATP[beta S]) isomers did not induce dissociation. The hydrolytic activity of the outer arm dynein from sea urchin sperm flagella towards these analogs was similar to that of 22S dynein. The ratios of Vmax (nmol.mg protein-1.min-1)/apparent Km (microM) of this dynein were 400-720, 53, 9.7, 0.62 and 0.028 for ATP, ATP[alpha S] (Sp or Rp), ATP[beta S] (Sp or Rp), respectively, in the presence of Mg2+ as the supporting cation. This dynein exhibited the same stereospecificity at beta phosphate as the 22S dynein or myosin. The detergent models of Tetrahymena or sea urchin spermatozoa were reactivated only by ATP or (Sp)ATP[alpha S] while other analogs were ineffective. The maximal beat frequency of the cilia or flagella reactivated by (Sp)ATP[alpha S] was one-quarter to one-half of that produced by ATP reactivation.     

Autoregulation of kinase dephosphorylation by ATP binding in AGC protein kinases

AGC kinases, including the three Akt (protein kinase B) isoforms, protein kinase A (PKA) and all protein kinase C (PKC) isoforms, require activation loop phosphorylation (threonine 308 in Akt1) as well as phosphorylation of a C-terminal residue (serine 473 in Akt1) for catalytic activity and phosphorylation of downstream targets. Conversely, phosphatases reverse these phosphorylations. Virtually all cellular processes are affected by AGC kinases, a circumstance that has led to intense scrutiny of the molecular mechanisms that regulate phosphorylation of these kinases. Here, we review a new layer of control of phosphorylation in Akt, PKA and PKC pointing to ATP binding pocket occupancy as a means to decelerate dephosphorylation of these and, potentially, other kinases. This additional level of kinase regulation opens the door to search for new functional motifs for the rational design of non- ATP-competitive kinase inhibitors that discriminate within and between protein kinase families.     

Autoregulation of kinase dephosphorylation by ATP binding to AGC protein kinases

AGC kinases, including the three Akt (protein kinase B) isoforms, protein kinase A (PKA) and all protein kinase C (PKC) isoforms, require activation loop phosphorylation (threonine 308 in Akt1) as well as phosphorylation of a C-terminal residue (serine 473 in Akt1) for catalytic activity and phosphorylation of downstream targets. Conversely, phosphatases reverse these phosphorylations. Virtually all cellular processes are affected by AGC kinases, a circumstance that has led to intense scrutiny of the molecular mechanisms that regulate phosphorylation of these kinases. Here, we review a new layer of control of phosphorylation in Akt, PKA and PKC pointing to ATP binding pocket occupancy as a means to decelerate dephosphorylation of these and, potentially, other kinases. This additional level of kinase regulation opens the door to search for new functional motifs for the rational design of non-ATP-competitive kinase inhibitors that discriminate within and between protein kinase families.Key words: inhibitors hijacking kinase activation, activation loop phosphorylation, dephosphorylation, phosphatase resistance, PKA, PKB, PKC     

ATP analogs substituted on the 2-position as substrates for dynein ATPase activity

In contrast to previously studied ATP analogs, the two-substituted ATP analogs, 2-N3 ATP and 2-Cl ATP were good substrates for dynein ATPase. The Vmax for hydrolysis of both analogs was significantly higher than for ATP and the Km for both analogs was comparable to ATP. The higher hydrolytic rate for the analogs might be explained by a faster dissociation rate of the diphosphate product. This interpretation is supported by measurements of the dissociation rate of the inhibitor, vanadate. The estimate dissociation rate of vanadate with the analogs as substrate is approx. 2-fold higher than with ATP as substrate. These data together with previous studies on a variety of ATP analogs suggest that the 6-amino group on adenine is important for recognition by dynein and that the anti-conformation of the adenine, favored by 2-substituents, is the favored conformation of the nucleotide.     

Site specific antibodies directed to the ATP binding region of some kinases

Polyclonal antibodies raised in rabbits to ATP-requiring enzymes such as 3-phosphoglycerate kinase show cross-reactivity against other unrelated kinases. Our results show that rabbit polyclonal antiserum possesses antibodies that recognize an antigenic site at the ATP binding region of kinases. A classical immunotitration curve was obtained when hexokinase was titrated against anti-myokinase IgG. The immunoinhibitions was reversed in the presence of small concentration of ATP. This cross-reactivity between site specific antibody and unrelated kinase demonstrates the existence of an antigenic site around the ATP binding region. Our proposal of the existence of a common antigenic determinant in the ATP binding region is in agreement with the finding of a common structural domain that binds ATP.     

Thiophosphate analogs of ADP and ATP as substrates in partial reactions of energy conversion in chloroplasts

Thiophosphate analogs of ADP and ATP have been employed in partial reactions of photosynthetic energy conversion in chloroplasts. Substitution of oxygen by sulfur at the α-phosphate yields a pair of diastereomers (ADPαS, ATPαS, A and B forms). Two diastereomeric compounds are also obtained by replacement of oxygen by sulfur in the β-phosphate group of ATP (ATPβS, A and B form) (Eckstein, F. and Goody, R.S. (1976) Biochemistry 15, 1685–1691).The A form of ADPαS is phosphorylated by chloroplasts with a K_m comparable to that of ADP but with a lower V. The B form of ADPαS as well as ADPβS is not a substrate in photophosphorylation and only weakly competes with ADP.The A forms of ADPαS and ATPαS strongly compete with ADP for the tight nucleotide binding site of CF₁ in the light-induced exchange reaction, whereas the B forms display a much smaller competitive effect. The efficiencies of ADPβS and the A isomer of ATPβS are intermediate, and the B form of ATPβS is a weaker competitor.The A forms of ATPαS and ATPβS are hydrolyzed by light-triggered ATPase, whereas the B forms are not. The efficiency of the A isomer of ATPαS is comparable to that of normal ATP, and the A form of ATPβS is cleaved at a lower rate. In trypsin-activated Ca²⁺-dependent ATPase the A form of ATPαS is the only thiophosphate analog to be hydrolyzed.The results indicate a stereospecific interaction of ADP and ATP at the catalytic sites as well as the tight nucleotide binding site of coupling ATPase of chloroplasts.     

Crystal structure of bacteriophage T4 deoxynucleotide kinase with its substrates dGMP and ATP.

NMP kinases catalyse the phosphorylation of the canonical nucleotides to the corresponding diphosphates using ATP as a phosphate donor. Bacteriophage T4 deoxynucleotide kinase (DNK) is the only member of this family of enzymes that recognizes three structurally dissimilar nucleotides: dGMP, dTMP and 5-hydroxymethyl-dCMP while excluding dCMP and dAMP. The crystal structure of DNK with its substrate dGMP has been determined at 2.0 A resolution by single isomorphous replacement. The structure of the ternary complex with dGMP and ATP has been determined at 2.2 A resolution. The polypeptide chain of DNK is folded into two domains of equal size, one of which resembles the mononucleotide binding motif with the glycine-rich P-loop. The second domain, consisting of five alpha-helices, forms the NMP binding pocket. A hinge connection between the domains allows for large movements upon substrate binding which are not restricted by dimerization of the enzyme. The mechanism of active centre formation via domain closure is described. Comparison with other P-loop-containing proteins indicates an induced-fit mode of NTP binding. Protein-substrate interactions observed at the NMP and NTP sites provide the basis for understanding the principles of nucleotide discrimination.     

Inhibition of the phosphorylase kinase activity by ATP analogs and their binding to the enzyme subunits

The interaction between phosphorylase kinase (EC 2.7.1.38), isolated from rabbit skeletal muscles, and the ATP analogs with the modified triphosphate fragment: adenosine-5'-chloromethane pyrophosphonate (1), adenosine-5'-chloroethyl phosphate (2), adenosine-5'-bromethane pyrophosphonate (3), adenosine-5'-bromoethane phosphonate (4), adenosine-5'-chloroacetylaminomethane phosphonate (5), adenosine-5'-chloroacetylaminomethane pyrophosphonate (6) and adenosine-5'-chloromethane phosphonate (7), was studied. The compounds 1, 2 and 3 irreversibly inhibit the enzyme activity. In the presence of ATP the rate of inactivation is decreased. The radioactive compounds 1, 2 and 3 are stoicheometrically incorporated into the beta- and gamma-subunits of phosphorylase kinase. A correlation is shown to exist between the degree of the beta-subunit modification by compound 1 and the enzyme inactivation. The compounds 4, 5 and 6 inhibit the enzyme reversibly: in the presence of ATP complete protection of the enzyme activity is observed. The compound 7 does not affect the kinase activity; however, it binds itself to the beta-subunit of the enzyme. The binding of analogs 1 and 7 to the beta-subunit occurs at different sites. The data obtained are indicative of the catalytic role of the beta-subunit of phosphorylase kinase.     

Screening of protein kinase inhibitors and knockdown experiments identified four kinases that affect mitochondrial ATP synthesis activity

Mitochondrial ATP synthase, a major ATP supplier in respiring cells, should be regulated in amount and in activity to respond to the varying demands of cells for ATP. We screened 80 protein kinase inhibitors and found that HeLa cells treated with four inhibitors exhibited reduced mitochondrial ATP synthesis activity. Consistently, knockdown of their target kinases (PKA, PKCδ, CaMKII and smMLCK) resulted in a decrease in mitochondrial ATP synthesis activity. Among them, mitochondria of smMLCK-knockdown cells contained only a small amount of ATP synthase, while the α- and β-subunits of ATP synthase were produced normally, suggesting that smMLCK affects assembly (or decay) of ATP synthase.     

Identifying specific matrix metalloproteinase-2-inhibiting peptides through phage display-based subtractive screening

Gelatinases A and B, which are members of the matrix metalloproteinase (MMP) family, play essential roles in cancer development and metastasis, as they can break down basal membranes. Therefore, the determination and inhibition of gelatinases is essential for cancer treatment. Peptides that can specifically block each gelatinase may, therefore, be useful for cancer treatment. In this study, subtractive panning was carried out using a 12-mer peptide library to identify peptides that block gelatinase A activity (MMP-2), which is a key pharmacological target. Using this method, 17 unique peptide sequences were determined. MMP-2 inhibition by these peptides was evaluated through zymogram analyses, which revealed that four peptides inhibited MMP-2 activity by at least 65%. These four peptides were synthesized and used for in vitro wound healing using human umbilical vein endothelial cells, and two peptides, AOMP12 and AOMP29, were found to inhibit wound healing by 40%. These peptides are, thus, potential candidates for MMP-2 inhibition for cancer treatment. Furthermore, our findings suggest that our substractive biopanning screening method is a suitable strategy for identifying peptides that selectively inhibit MMP-2.     

Inhibition of adenosine and thymidylate kinases by bisubstrate analogs

Potential bisubstrate analogs, in which the 5'-hydroxyl group of adenosine was joined to the phosphoryl group acceptor by polyphosphoryl bridges of varying length (ApnX, where n is the number of phosphoryl groups and X is the nucleoside moiety of the acceptor), were tested as inhibitors of human liver adenosine kinase and of thymidylate kinase from peripheral blast cells of patients with acute myelocytic leukemia. Adenosine kinase was most strongly inhibited by P1,P4-(diadenosine 5')-tetraphosphate (Kd = 30 nM) and P1,P5-(diadenosine 5')-pentaphosphate (Kd = 73 nM). Thymidylate kinase was most strongly inhibited by P1-(adenosine 5')-P5-(thymidine 5')-pentaphosphate (Kd = 120 nM) and by P1(adenosine 5')-P6-(thymidine 5')-hexaphosphate (Kd = 43 nM). In these enzymes, as in adenylate and thymidylate kinases, strongest inhibition was achieved in compounds containing one or two more phosphoryl groups than the substrates combined. These results support the view that nucleoside and nucleotide kinases mediate direct transfer of phosphoryl groups from ATP to acceptors, rather than acting by a double displacement mechanism.     

Phosphorylation of Rho-associated kinase (Rho-kinase/ROCK/ROK) substrates by protein kinases A and C

Rho-associated kinase (Rho-kinase/ROCK/ROK) is a serine/threonine kinase and plays an important role in various cellular functions. The cAMP-dependent protein kinase (protein kinase A/PKA) and protein kinase C (PKC) are also serine/threonine kinases, and directly and/or indirectly take part in the signal transduction pathways of Rho-kinase. They have similar phosphorylation site motifs, RXXS/T and RXS/T. The purpose of this study was to identify whether sites phosphorylated by Rho-kinase could be targets for PKA and PKC and to find peptide substrates that are specific to Rho-kinase, i.e., with no phosphorylation by PKA and PKC. A total of 18 substrates for Rho-kinase were tested for phosphorylation by PKA and PKC. Twelve of these sites were easily phosphorylated. These results mean that Rho-kinase substrates can be good substrates for PKA and/or PKC. On the other hand, six Rho-kinase substrates showing no or very low phosphorylation efficiency (<20%) for PKA and PKC were identified. Kinetic parameters (K(m) and k(cat)) showed that two of these peptides could be useful as substrates specific to Rho-kinase phosphorylation.     

Calcium signaling through protein kinases. The Arabidopsis calcium-dependent protein kinase gene family

In plants, numerous Ca(2+)-stimulated protein kinase activities occur through calcium-dependent protein kinases (CDPKs). These novel calcium sensors are likely to be crucial mediators of responses to diverse endogenous and environmental cues. However, the precise biological function(s) of most CDPKs remains elusive. The Arabidopsis genome is predicted to encode 34 different CDPKs. In this Update, we analyze the Arabidopsis CDPK gene family and review the expression, regulation, and possible functions of plant CDPKs. By combining emerging cellular and genomic technologies with genetic and biochemical approaches, the characterization of Arabidopsis CDPKs provides a valuable opportunity to understand the plant calcium-signaling network.     

Achieving specific RNA cleavage activity by an inactive splicing endonuclease subunit through engineered oligomerization

Protein-protein interaction is a common strategy exploited by enzymes to control substrate specificity and catalytic activities. RNA endonucleases, which are involved in many RNA processing and regulation processes, are prime examples of this. How the activities of RNA endonucleases are tightly controlled such that they act on specific RNA is of general interest. We demonstrate here that an inactive RNA splicing endonuclease subunit can be switched "on" solely by oligomerization. Furthermore, we show that the mode of assembly correlates with different RNA specificities. The recently identified splicing endonuclease homolog from Sulfolobus solfataricus, despite possessing all of the putatively catalytic residues, has no detectable RNA cleavage activity on its own but is active upon mixing with its structural subunit. Guided by the previously determined three-dimensional structure of the catalytic subunit, we altered its sequence such that it could potentially self-assemble thereby enabling its catalytic activity. We present the evidence for the specific RNA cleavage activity of the engineered catalytic subunit and for its formation of a functional tetramer. We also identify a higher order oligomer species that possesses distinct RNA cleavage specificity from that of previously characterized RNA splicing endonucleases.     

Isolation and characterization of genetically engineered gallidermin and epidermin analogs.

Gallidermin (Gdm) and epidermin (Epi) are highly homologous tetracyclic polypeptide antibiotics that are ribosomally synthesized by a Staphylococcus gallinarum strain and a Staphylococcus epidermidis strain, respectively. These antibiotics are secreted into media and are distinguished by the presence of the unusual amino acids lanthionine, 3-methyllanthionine, didehydrobutyrine, and S-(2-aminovinyl)-D-cysteine, which are formed by posttranslational modification. To study the substrate specificities of the modifying enzymes and to obtain variants that exhibit altered or new biological activities, we changed certain amino acids by performing site-specific mutagenesis with the Gdm and Epi structural genes (gdmA and epiA, respectively). S. epidermidis Tü3298/EMS6, an epiA mutant of the Epi-producing strain, was used as the expression host. This mutant synthesized Epi, Gdm, or analogs of these antibiotics when the appropriate genes were introduced on a plasmid. No Epi or Gdm analogs were isolated from the supernatant when (i) hydroxyamino acids involved in thioether amino acid formation were replaced by nonhydroxyamino acids (S3N and S19A); (ii) C residues involved in thioether bridging were deleted (delta C21, C22 and delta C22); or (iii) a ring amino acid was replaced by an amino acid having a completely different character (G10E and Y20G). A strong decrease in production was observed when S residues involved in thioether amino acid formation were replaced by T residues (S16T and S19T). A number of conservative changes at positions 6, 12, and 14 on the Gdm backbone were tolerated and led to analogs that had altered biological properties, such as enhanced antimicrobial activity (L6V) or a remarkable resistance to proteolytic degradation (A12L and Dhb14P). The T14S substitution led to simultaneous production of two Gdm species formed by incomplete posttranslational modification (dehydration) of the S-14 residue. The fully modified Dhb14Dha analog exhibited antimicrobial activity similar to that of Gdm, whereas the Dhb14S analog was less active. Both peptides were more sensitive to tryptic cleavage than Gdm was.