Single Amino Acids Determine Specificity of Binding of Protein Kinase A Regulatory Subunits by Protein Kinase A Anchoring Proteins
1999; Elsevier BV; Volume: 274; Issue: 41 Linguagem: Inglês
10.1074/jbc.274.41.29057
ISSN1083-351X
Autores Tópico(s)Protein Kinase Regulation and GTPase Signaling
ResumoCyclic AMP-dependent protein kinase is tethered to protein kinase A anchoring proteins (AKAPs) through regulatory subunits (R) by RIα-specific, RIIα-specific, or RIα/RIIα dual-specific binding. Ala- and Val-scanning mutagenesis determined that hydrophobic amino acids at three homologous positions are required for binding of RIα to FSC1/AKAP82 domain B and RIIα to AKAP Ht31. A mutation at the middle position reversed the binding specificity of both AKAPs, and mutations at this same position of the dual-specific domain A of FSC1/AKAP82 converted it into either an RIα or RIIα binding domain. This suggests that hydrophobic amino acids at three conserved positions within the primary sequence and an amphipathic helix of AKAPs are required for cyclic AMP-dependent protein kinase binding, with the size of the aliphatic side chain at the middle position determining RIα or RIIα binding specificity. Cyclic AMP-dependent protein kinase is tethered to protein kinase A anchoring proteins (AKAPs) through regulatory subunits (R) by RIα-specific, RIIα-specific, or RIα/RIIα dual-specific binding. Ala- and Val-scanning mutagenesis determined that hydrophobic amino acids at three homologous positions are required for binding of RIα to FSC1/AKAP82 domain B and RIIα to AKAP Ht31. A mutation at the middle position reversed the binding specificity of both AKAPs, and mutations at this same position of the dual-specific domain A of FSC1/AKAP82 converted it into either an RIα or RIIα binding domain. This suggests that hydrophobic amino acids at three conserved positions within the primary sequence and an amphipathic helix of AKAPs are required for cyclic AMP-dependent protein kinase binding, with the size of the aliphatic side chain at the middle position determining RIα or RIIα binding specificity. cAMP-dependent protein kinase regulatory regulatory subunit type I regulatory subunit type II A-kinase anchoring protein glutathione S- transferase An important problem for the cell is how to produce a localized response to a freely diffusable signal transduction product. This is particularly true when the response involves activating one of the many protein kinases, such as a member of the cAMP-dependent protein kinase (PKA)1 family. The PKAs usually exist as inactive tetramers containing a regulatory (R) subunit dimer and two catalytic subunits, and genes encoding four R subunits and three catalytic subunits have been identified (1Takio K. Smith S.B. Krebs E.G. Walsh K.A. Titani K. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 2544-2548Crossref PubMed Scopus (83) Google Scholar, 2Lee D.C. Carmichael D.F. Krebs E.G. McKnight G.S. Proc. Natl. Acad. Sci. U. S. A. 1983; 80: 3608-3612Crossref PubMed Scopus (132) Google Scholar, 3Jahnsen T. Hedin L. Kidd V.J. Beattie W.G. Lohmann S.M. Walter U. Durica J. Schulz T.Z. Schiltz E. Browner M. Lawrence C.B. Goldman D. Ratoosh S.L. Richards J.S. J. Biol. Chem. 1986; 261: 12352-12361Abstract Full Text PDF PubMed Google Scholar, 4Clegg C.H. Cadd G.G. McKnight G.S. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 3703-3707Crossref PubMed Scopus (167) Google Scholar). The type I (RI) α and type II (RII) α subunits are distributed ubiquitously, whereas RIβ and RIIβ are present mainly in brain. Many hormones and other signals act through receptors to generate cAMP that binds to the R subunits of PKA, thereby releasing and activating the catalytic subunits to phosphorylate proteins. The way cells produce a localized response to cAMP has been determined for some PKAs. They are tethered through their RII dimer to protein kinase A anchoring proteins (AKAPs) that place them in close proximity to specific organelles or cytoskeletal components (5Scott J.D. Stofko R.E. McDonald J.R. Comer J.D. Vitalis E.A. Mangili J.A. J. Biol. Chem. 1990; 265: 21561-21566Abstract Full Text PDF PubMed Google Scholar, 6Bregman D.B. Hirsch A.H. Rubin C.S. J. Biol. Chem. 1991; 266: 7207-7213Abstract Full Text PDF PubMed Google Scholar). More than 20 AKAPs have been identified that bind RII in overlay assays and localize to different subcellular compartments. The 10–14 residues comprising RII anchoring domains of AKAPs vary substantially in primary sequence, but secondary structure predictions indicate they are likely to form an amphipathic helix (7Carr D.W. Stofko-Hahn R.E. Fraser I.D. Bishop S.M. Acott T.S. Brennan R.G. Scott J.D. J. Biol. Chem. 1991; 266: 14188-14192Abstract Full Text PDF PubMed Google Scholar, 8Carr D.W. Hausken Z.E. Fraser I.D. Stofko-Hahn R.E. Scott J.D. J. Biol. Chem. 1992; 267: 13376-13382Abstract Full Text PDF PubMed Google Scholar). Analysis of the domain on AKAP75 by Ala-scanning mutagenesis suggested that hydrophobic amino acids with a long aliphatic side chain (e.g. Val, Leu, or Ile) participate in the binding to RII (9Glantz S.B. Li Y. Rubin C.S. J. Biol. Chem. 1993; 268: 12796-12804Abstract Full Text PDF PubMed Google Scholar). NMR studies of the RIIα dimerization/docking domain suggested that a hydrophobic surface associates with the amphipathic helix on AKAP Ht31 (10Newlon M.G. Roy M. Morikis D. Hausken Z.E. Coghlan V. Scott J.D. Jennings P.A. Nat. Struct. Biol. 1999; 6: 222-227Crossref PubMed Scopus (184) Google Scholar). These and other findings (11Hausken Z.E. Coghlan V.M. Hastings C.A. Reimann E.M. Scott J.D. J. Biol. Chem. 1994; 269: 24245-24251Abstract Full Text PDF PubMed Google Scholar, 12Li Y. Rubin C.S. J. Biol. Chem. 1995; 270: 1935-1944Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar, 13Hausken Z.E. Dell'Acqua M.L. Coghlan V.M. Scott J.D. J. Biol. Chem. 1996; 271: 29016-29022Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar) strongly suggest that hydrophobic interactions have a key role in the interactions between AKAPs and RII dimers. Overlay assays failed to detect binding by other PKA subunits to AKAPs, but indirect evidence suggested that RIα could bind to AKAPs. Comparison of the dimerization/docking domains of RIα and RII indicated striking similarities (14Newlon M.G. Roy M. Hausken Z.E. Scott J.D. Jennings P.A. J. Biol. Chem. 1997; 272: 23637-23644Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar, 15León D.A. Herberg F.W. Banky P. Taylor S.S. J. Biol. Chem. 1997; 272: 28431-28437Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar, 16Banky P. Huang L.J.-S. Taylor S.S. J. Biol. Chem. 1998; 273: 35048-35055Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar), and PKA appeared to be localized by RIα in skeletal muscle of RIIα knockout mice (17Burton K.A. Johnson B.D. Hausken Z.E. Westenbroek R.E. Idzerda R.L. Scheuer T. Scott J.D. Catterall W.A. McKnight G.S. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 11067-11072Crossref PubMed Scopus (118) Google Scholar). Other studies suggested that RI could associate with the plasma membrane (18Rubin C.S. Erlichman J. Rosen O.M. J. Biol. Chem. 1972; 247: 6135-6139Abstract Full Text PDF PubMed Google Scholar), neuromuscular junctions (19Imaizumi-Scherrer T. Faust D.M. Bénichou J.C. Hellio R. Weiss M.C. J. Cell Biol. 1996; 134: 1241-1254Crossref PubMed Scopus (44) Google Scholar), and the sperm flagellum (20Moos J. Peknicová J. Geussová G. Philimonenko V. Hozák P. Mol. Reprod. Dev. 1998; 50: 79-85Crossref PubMed Scopus (18) Google Scholar). In addition, yeast two-hybrid assays indicate that D-AKAP1/S-AKAP84 (21Huang L.J. Durick K. Weiner J.A. Chun J. Taylor S.S. J. Biol. Chem. 1997; 272: 8057-8064Abstract Full Text Full Text PDF PubMed Scopus (257) Google Scholar, 22Lin R.-Y. Moss S.B. Rubin C.S. J. Biol. Chem. 1995; 270: 27804-27811Abstract Full Text Full Text PDF PubMed Scopus (161) Google Scholar) and D-AKAP2 (23Huang L.J. Durick K. Weiner J.A. Chun J. Taylor S.S. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 11184-11189Crossref PubMed Scopus (202) Google Scholar) interact with both RIα and RIIα. The first direct evidence of RI binding to an AKAP was found in recent studies that identified RIα-specific and RIα/RIIα dual specificity PKA anchoring domains on FSC1/AKAP82 (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). Deletion mutagenesis and yeast two-hybrid assays were used to map the RIα-specific domain to a 10-amino acid sequence likely to form an amphipathic helix. It has little significant sequence homology to RII anchoring domains and a low content of hydrophobic amino acids with a long aliphatic side chain. The binding of RIα was not affected by substitution of Ser for Val339, the only hydrophobic amino acid with a long aliphatic side chain in the RIα-specific domain. The sequences flanking the 10 amino acids of this domain were not essential for RIα binding in the yeast two-hybrid system, strongly suggesting that the domain contains all of the information necessary to confer RIα-specific binding (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). The RIα/RIIα dual specificity domain mapped to a 14-amino acid sequence that contains 5 hydrophobic amino acids with a long aliphatic side chain, residues which frequently appear in RII anchoring domains (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). The demonstration that RIα could tether PKA to RIα-specific or RIα/RIIα dual specificity domains raised important questions about the primary and secondary structural features that determine the nature and specificity of PKA anchoring domains on AKAPs. The present studies used scanning mutagenesis of the RIα-specific domain of FSC1/AKAP82 (domain B) and the putative RII-specific domain of Ht31 to demonstrate that single amino acid residues in these domains determine RI and RII binding specificity. Furthermore, point mutations introduced into the RIα/RIIα dual specificity domain of FSC1/AKAP82 (domain A) converted it into an RI or RII preferential binding domain. These results lead us to propose that the degree and specificity of PKA binding to anchoring domains on AKAPs are determined by the size of the aliphatic side chain at a consensus position within the amphipathic helix. The Ht31(GenBankTM accession number, M90360) fragment (residues 418–540) containing the RII anchoring domain was amplified by polymerase chain reaction using a 5′-primer containing aBamHI site (CCGGGATCCATGGGTGACGCTGAGGAAGCCCA), a 3′-primer containing a SalI site (CCGGTCGACTCTCTAGTCCTTTAGTGAGAGGAC), and the HT31 cDNA cloned in pET 11d (gift of Dr. Daniel W. Carr, Veterans Affairs Medical Center, Portland, OR) as a template (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). The reaction products were digested withBamHI and SalI, ligated into plasmid pGEX-4T-1 (Amersham Pharmacia Biotech) digested with BamHI andSalI, and then transformed in Escherichia coliDH5α-competent cells (Life Technologies, Inc.). This plasmid and pGEX plasmids containing FSC1 coding sequence residues 202–334 or 237–361 (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar) were used as templates to introduce point mutations using the QuikChangeTM site-directed mutagenesis kit (Stratagene). Numbering of the amino acid sequence of FSC1 was that of Fulcheret al. (25Fulcher K.D. Mori C. Welch J.E. O'Brien D.A. Klapper D.G. Eddy E.M. Biol. Reprod. 1995; 52: 41-49Crossref PubMed Scopus (83) Google Scholar). Sequencing of the targeted sites was carried out using a dRhodamine Terminator Cycle Sequencing kit (PE Biosystems) to verify that the correct mutations occurred. The pull-down experiments using glutathione-Sepharose (Amersham Pharmacia Biotech) and the glutathione S-transferase (GST) expression tag system were performed as described previously (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar) with the following minor changes. Testis crude extracts corresponding to 1.5 mg of protein were used for incubation with GST-FSC1 or GST-Ht31 fusion proteins immobilized on resin. Each lane of the gels received 0.5 mg of protein from the testis extracts that bound to GST-FSC1 or GST-Ht31, whereas control lanes received 5 μg of total testis extract protein. Regulatory subunits of PKA were detected by Western blotting using a monoclonal antibody against mouse RIα (1:250 dilution; Transduction Laboratories, catalog number P19920) or a rabbit antiserum to mouse RIIα (1:1000 dilution; Santa Cruz Biotechnology, catalog number sc-909). Secondary antibodies were horseradish peroxidase-conjugated goat antiserum to mouse IgG (1:30000 dilution; Sigma) or goat antiserum to rabbit IgG (1:30000 dilution; Santa Cruz Biotechnology). All other procedures for Western blotting and chemiluminescence were performed as described previously (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). The Ht31 fragment (residue 418–540) containing the RII anchoring domain was amplified as described above except the 5′-primer contained an EcoRI site (CCGGAATTCATGGGTGACGCTGAGGAAGCCCAAAT) instead of a BamHI site. The reaction products were digested with EcoRI andSalI, ligated into plasmid pAS2-1 (CLONTECH) digested with EcoRI andSalI, and then transformed in E. coliDH5α-competent cells (Life Technologies, Inc.). The yeast two-hybrid system was used for analysis of interactions between an Ht31 fragment containing the RII anchoring domain and a full-length RIα sequence (FC9) cloned into pGAD10 (CLONTECH) as described before (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). Transactivation of his3 and lacZ was used for the detection of protein-protein interaction. RII binding is believed to occur by forming hydrophobic interactions (15León D.A. Herberg F.W. Banky P. Taylor S.S. J. Biol. Chem. 1997; 272: 28431-28437Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar), and domain B of FSC1/AKAP82 has similarities to RII binding domains of other AKAPs. However, the only hydrophobic amino acid present with a long aliphatic side chain (Val339) is not necessary for RIα binding (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). We hypothesized that other hydrophobic amino acids in domain B (Ala, Tyr, and Met) are used to form the amphipathic helix and are involved in RIα binding. This was tested using in vitro pull-down assays in conjunction with Val-scanning and Ala-scanning mutagenesis. These substitutions were made for each hydrophobic amino acid in domain B except Met344, where Ile was introduced. GST-FSC1 fusion proteins (residues 237–361) containing each mutation were immobilized on glutathione-Sepharose resin and incubated with crude extracts of testis to determine if they bound RIα or RIIα. RIα binding to domain B was lost or reduced with Y335V, A336V, and A340V mutations but retained with M343V/M344I double mutations (Fig.1 a). Alanine-scanning mutagenesis done in parallel resulted in loss of RIα binding with Y335A and M344A mutations but not with V339A or M343A mutations. In addition, it was found that the A340V mutation resulted in RIIα binding (Fig. 1 a). These results have three significant implications. First, the same amino acid sequence can allow both RIα and RIIα binding. Second, one of the four hydrophobic amino acids important for RIα binding (Ala340) determines the specificity for RIα binding. Third, the size of the hydrophobic side chains is a key factor affecting the binding and specificity of RIα and RIIα subunits of PKA. The RII anchoring domain of Ht31 (residues 494–507) contains six amino acids with a long aliphatic side chain. We supposed that some of these amino acids function in RIIα-specific binding as do those of AKAP75 (9Glantz S.B. Li Y. Rubin C.S. J. Biol. Chem. 1993; 268: 12796-12804Abstract Full Text PDF PubMed Google Scholar). We used Ala-scanning mutagenesis and in vitro pull-down assays to determine which of these residues are essential for RII binding and if any mutations result in RI binding (Fig. 1 b). The I502A, V503A, and I507A mutations abolished RIIα binding. Val-scanning mutagenesis demonstrated that Ala499 also is important for RIIα binding. In addition, the V503A mutation resulted in RIα binding, whereas binding was detected with the L494A, A498V, and V506A mutants after longer exposure to x-ray film (data not shown). RIα binding to the intact Ht31 domain (Fig.2 b) was verified using a yeast two-hybrid assay (data not shown). These results indicate that significant differences occur between the amino acids required for isoform-specific binding on domain B and Ht31. However, replacement of Ala340 with Val resulted in RIIα binding to domain B, and conversely the replacement of Val503 with Ala resulted in RIα binding to Ht31, indicating the importance of individual residues in determining isoform-specific binding. The next set of studies examined the basis for differences between the amino acids required for isoform-specific binding on domain B and Ht31. Point mutagenesis andin vitro pull-down assays were used to determine if other amino acids could be substituted at the sites critical for isoform-specific binding. The binding of RIα to domain B was reduced and RIIα binding initiated when Ala340 was replaced with Val or Ile (Fig. 2 a). The substitution of Leu resulted in loss of binding of RIα and slight RIIα binding. The replacement of Ala with Phe, a hydrophobic amino acid containing an aromatic side chain, or with Lys, a hydrophilic amino acid with a bulky side chain, abolished RIα binding and did not lead to RIIα binding. These results indicate that RIIα binding to domain B occurs when hydrophobic amino acids with a long aliphatic side chain (e.g. Val, Ile, or Leu) are positioned at the critical location in the amphipathic helix. Amino acid substitution was also used to analyze the critical site (Val503) for RII binding specificity on Ht31 (Fig.2 b). The only substitution for this amino acid that increased RIα binding was Ala (V503A), whereas substitution of Ser (V503S) abolished both RIIα binding and the weaker RIα binding. This suggests that RIα binding is determined by a hydrophobic amino acid with a small side chain (e.g. Ala) at this site. Substitution of Ile (V503I) did not have a major effect on RIα or RIIα binding, but substitution of Leu (V503L) substantially reduced RIIα binding. These findings and the results of mutagenesis studies on Ala340 of domain B indicate that RIIα binding occurs best when Val and Ile are present in the critical sites. These results lead us to hypothesize that binding specificity is determined by the size of the aliphatic side chain at the critical amino acid, with a small side chain specifying RIα binding and a large side chain specifying RIIα binding. These results allowed us to determine the consensus amino acid sequence (Fig. 3 a) and predicted secondary structure (Fig. 3 b) of the domain B and Ht31-anchoring sites for RIα and RIIα. The consensus primary sequence (Fig. 3 a) was determined by aligning Ala340 of domain B and Val503 of Ht31 at position 6 (shown by circles). This resulted in positions 2 and 10 containing two of the three hydrophobic amino acid residues (shown by squares) important for RIα or RIIα binding on domain B or Ht31. Both anchoring domains contain Ala at position 2, whereas position 10 contains a hydrophobic amino acid with a bulk side chain (Met or Ile). The predicted secondary structure was determined using a computer software program (DNASISTM). A helical wheel projection of the amino acids at positions 1–10 on domain B and Ht31 was used to display their relative positions within an amphipathic helix (Fig.3 b). For an α-helix viewed from the amino end, residue positions 2, 6, and 10 are seen to form a cluster. Alanine at position 2 and the hydrophobic amino acid with a bulk side chain at position 10 embrace the key residue at position 6 that determines isoform-specific binding. An additional hydrophobic amino acid with a bulk side chain is located peripherally, at position 1 for domain B and position 5 for Ht31. This suggests that positions 2, 6, and 10 on the amphipathic helix form the core of the common RIα and RIIα binding domain, whereas the additional hydrophobic amino acids with a bulk side chain at variable peripheral positions also are required for regulatory subunit binding to each AKAP domain. Domain A of FSC1/AKAP82 binds both RIα and RIIα (24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar,26Visconti P.E. Johnson L.R. Oyaski M. Fornes M. Moss S.B. Gerton G.L. Kopf G.S. Dev. Biol. 1997; 192: 351-363Crossref PubMed Scopus (175) Google Scholar). This domain was analyzed next to determine the features responsible for the RIα/RIIα dual binding capabilities and to relate them to the features of RIα-specific domain B and the RIIα-specific Ht31 domain. A fragment of FSC1/AKAP82 (residues 202–334) containing domain A was analyzed by a combination of point mutagenesis and in vitro pull-down assays to identify the key amino acids that participate in the dual binding ability (Fig.4). Domain A contains an Ala, but it is located at the C terminus of the putative RII binding domain (26Visconti P.E. Johnson L.R. Oyaski M. Fornes M. Moss S.B. Gerton G.L. Kopf G.S. Dev. Biol. 1997; 192: 351-363Crossref PubMed Scopus (175) Google Scholar) and seems unlikely to determine binding specificity (Fig. 3 a). Therefore Ala-scanning mutagenesis was used to modify other selected hydrophobic amino acid residues in domain A. As shown in Fig.4 a, the I229A mutation abolished RIIα binding and did not lead to an increase but rather a slight decrease in RIα binding (Fig.4 a). However, substitution of Ala for Val221 (V221A) dramatically increased both RIα and RIIα binding, whereas the Y220A and L227A mutations slightly increased binding by RIIα or both subunits, respectively. Substitution of Ser for Val221(V221S) produced a moderate increase in both RIα and RIIα binding (Fig. 4 b). These results indicate that when residue 221 is a hydrophobic amino acid with a small side chain, it enables binding of RI and RII subunits to domain A. This information was used to compare domain A with the binding sequences determined for domain B and Ht31 (Fig. 3 a). Val221 was placed at position 2 on the amphipathic helix (shown by a triangle), the position of the residue critical for RIα binding to domain B and RIIα binding to Ht31. This alignment placed Ile229 at position 10 (shown by asquare), where our results indicate that a hydrophobic amino acid with a bulk side chain is required for strong RI binding to domain B or RII binding to Ht31. The alignment also placed Ser225at position 6 (shown by a diamond, Fig. 3, a andb) where a hydrophobic amino acid (Ala, Val, Ile, or Leu) was required for isoform-specific binding to domain B or Ht31. Substitutions for Ser225 verified that this position is involved in determining isoform-specific binding (Fig. 4 b). Replacement of Ser225 with Ala (S225A) did not increase RIα binding, whereas replacement with Val (S225V) resulted in some RIIα binding. We replaced Val221 with Ala (V221A) in domain A to enhance RIα and RIIα binding prior to further analyzing isoform specificity. A double mutation that included substitution of Ala for Ser225 (V221A/S225A) reduced RIIα binding, but RIα binding was little changed compared with the V221A mutation (Fig.4 b). However, substitution of Val for Ser225(V221A/S225V) dramatically enhanced RIIα binding and abolished RIα binding induced by the V221A mutation. These results indicate that Ser225 is the key position for determining isoform-specific binding to domain A. They also demonstrate that replacement of Ser225 with Ala or Val converts the anchoring domain from one of dual specificity to an RIα- or RIIα-preferential anchoring domain, respectively. The AKAPs had been shown only to anchor PKAs by RII subunit dimers until an RIα-specific and a dual binding domain for RI and RII were identified recently (21Huang L.J. Durick K. Weiner J.A. Chun J. Taylor S.S. J. Biol. Chem. 1997; 272: 8057-8064Abstract Full Text Full Text PDF PubMed Scopus (257) Google Scholar, 23Huang L.J. Durick K. Weiner J.A. Chun J. Taylor S.S. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 11184-11189Crossref PubMed Scopus (202) Google Scholar, 24Miki K. Eddy E.M. J. Biol. Chem. 1998; 273: 34384-34390Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar). This added a new dimension to the role of AKAPs in the localization and function of different PKAs at specific subcellular sites. It also raised the question of the basis for the specificity of binding of different PKA isozymes. A prominent difference between the RI and RII anchoring domains is that the RI domain has a lower content of amino acids with a long aliphatic side chain (Val, Leu, or Ile). This led us in the present study to analyze isoform-specific binding by Val- and Ala-scanning mutagenesis. We found the following: (i) hydrophobic amino acids in the anchoring domains are essential for isoform-specific binding; (ii) two amino acids required for binding are located at consensus positions 2 and 10 of an amphipathic helix, and an amino acid required for isoform-specific binding is located at consensus position 6; (iii) the amino acid at position 2 has a small side chain, and the amino acid at position 10 has a hydrophobic amino acid with a bulk side chain; and (iv) binding specificity is determined by the size of the hydrophobic side chain at position 6, with amino acids bearing a small side chain necessary for RIα binding and those with a long aliphatic side chain necessary for RIIα binding (Fig.5). In addition, each of the AKAP anchoring domains analyzed required distinctive hydrophobic amino acids at other positions for PKA binding. The specific requirements at positions 2, 6, and 10 indicate that these three amino acids form the basis of regulatory subunit binding. We refer to this arrangement as the three amino acid rule. However, this arrangement is necessary but not sufficient for PKA binding, and additional hydrophobic amino acids located at other position(s) are required for subunit anchoring. The three amino acid rule is applicable to other AKAPs. RII anchoring was eliminated by substitution of Ala for Ile405 on AKAP75 and Ala for Ile248 on AKAPCE (9Glantz S.B. Li Y. Rubin C.S. J. Biol. Chem. 1993; 268: 12796-12804Abstract Full Text PDF PubMed Google Scholar, 27Angelo R. Rubin C.S. J. Biol. Chem. 1998; 273: 14633-14643Abstract Full Text Full Text PDF PubMed Scopus (92) Google Scholar). An alignment that places these residues at position 10 locates Ala at position 2 for both AKAPs. Furthermore, AKAP75 has a Val at position 6 as does Ht31, and AKAPCE has a Ser at position 6 as occurs in domain A. Vijayaraghavan et al. (28Vijayaraghavan S. Liberty G.A. Mohan J. Winfrey V.P. Olson G.E. Carr D.W. Mol. Endocrinol. 1999; 13: 705-717Crossref PubMed Scopus (0) Google Scholar) aligned AKAP110 with 14 other AKAPs and identified a consensus sequence that is consistent with the three amino acid rule. However, there are a few exceptions, including the presence of Val, a hydrophobic amino acid with a long aliphatic side chain at position 2 of domain A. Substitution of Ala at this position (V221A) dramatically increased RIα and RIIα binding to domain A (Fig. 4), as predicted by the three amino acid rule. Ht31 was considered to be an RII-specific AKAP (29Rosenmund C. Carr D.W. Bergeson S.E. Nilaver G. Scott J.D. Westbrook G.L. Nature. 1994; 368: 853-856Crossref PubMed Scopus (324) Google Scholar), but we found that Ht31 binds RIIα preferentially and RIα weakly. We have also observed that AKAP79 binds RIα as well as RIIα. 2K. Miki and E. M. Eddy, unpublished results. This suggests that if an anchoring domain has a hydrophobic amino acid with a long aliphatic side chain at position 6, it has RII-preferential, dual-specific binding (RII > RI). However, RIα and RIIα showed comparable binding to domain A. We refer to this type of dual-specific binding as "neutral specificity" (RI ≈ RII). RII-preferential, dual-specific binding may occur with D-AKAP1 and D-AKAP2, as occurs with Ht31. The neutral specificity of domain A appears to be due to the presence of Ser at position 6 instead of an Ala, Val, Ile, or Leu residue present at this position in other AKAPs. The structural features of the binding site on RI and RII dimers that associate with AKAPs have been examined recently by NMR and x-ray crystallography. RIα and RIIα have limited sequence homology, but both possess an α-helical region at the N terminus that is responsible for regulatory subunit dimer formation and for docking the dimer to AKAPs (14Newlon M.G. Roy M. Hausken Z.E. Scott J.D. Jennings P.A. J. Biol. Chem. 1997; 272: 23637-23644Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar, 15León D.A. Herberg F.W. Banky P. Taylor S.S. J. Biol. Chem. 1997; 272: 28431-28437Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar). An NMR study of the N-terminal region of RIIα (residues 1–44) indicated that a four-helix bundle dimerization motif with an extended hydrophobic face defined the AKAP-binding site (10Newlon M.G. Roy M. Morikis D. Hausken Z.E. Coghlan V. Scott J.D. Jennings P.A. Nat. Struct. Biol. 1999; 6: 222-227Crossref PubMed Scopus (184) Google Scholar). Studies using point mutagenesis of the dimerization/docking domain of RIα, RIIα, and RIIβ have indicated that hydrophobic amino acids with bulk side chains are required for binding to AKAPs without abolishing dimerization (11Hausken Z.E. Coghlan V.M. Hastings C.A. Reimann E.M. Scott J.D. J. Biol. Chem. 1994; 269: 24245-24251Abstract Full Text PDF PubMed Google Scholar, 12Li Y. Rubin C.S. J. Biol. Chem. 1995; 270: 1935-1944Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar,16Banky P. Huang L.J.-S. Taylor S.S. J. Biol. Chem. 1998; 273: 35048-35055Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). These results are in agreement with our data indicating that hydrophobic interactions are a crucial factor for association between AKAPs and PKA regulatory subunits. However, Newlon et al.(10Newlon M.G. Roy M. Morikis D. Hausken Z.E. Coghlan V. Scott J.D. Jennings P.A. Nat. Struct. Biol. 1999; 6: 222-227Crossref PubMed Scopus (184) Google Scholar) observed NMR chemical shift changes not only for hydrophobic but also for hydrophilic amino acid residues corresponding to the dimerization/docking domain of RIIα upon addition of a peptide containing the Ht31 anchoring domain. Although the chemical shift changes could result from structural changes upon binding of the Ht31 peptide, we cannot exclude the possibility that hydrophilic amino acids of AKAP anchoring domains have direct contact with regulatory subunits and modulate binding affinity and/or specificity. Protein-protein binding depends upon different molecular interactions, including ionic, hydrophilic, and hydrophobic interactions. Alanine-scanning mutagenesis is widely used to evaluate the function of intermolecular and intramolecular hydrophobic interactions (9Glantz S.B. Li Y. Rubin C.S. J. Biol. Chem. 1993; 268: 12796-12804Abstract Full Text PDF PubMed Google Scholar, 30Sattler M. Liang H. Nettesheim D. Meadows R.P. Harlan J.E. Eberstadt M. Yoon H.S. Shuker S.B. Chang B.S. Minn A.J. Thompson C.B. Fesik S.W. Science. 1997; 275: 983-986Crossref PubMed Scopus (1300) Google Scholar, 31Darimont B.D. Wagner R.L. Apriletti J.W. Stallcup M.R. Kushner P.J. Baxter J.D. Fletterick R.J. Yamamoto K.R. Genes Dev. 1998; 12: 3343-3356Crossref PubMed Scopus (832) Google Scholar). When protein interactions are disturbed by an Ala substitution, they are likely due to hydrophobic interactions. An unexpected finding of the present study was that Val-scanning mutagenesis revealed indispensable functions of Ala residues other than forming hydrophobic bonds. These results indicated that (i) protein interactions can be defined by the nature of the hydrophobic face, (ii) binding specificity can be determined by a single amino acid based on the size of the side chain, (iii) and Val-scanning mutagenesis is useful for identifying the function of a small hydrophobic side chain on an amphipathic helix. Hydrophobic protein interactions also occur between leucine zipper motifs, with the hydrophobic amino acids arrayed in an α-helix that forms a line of long aliphatic groups (32Landschulz W.H. Johnson P.F. McKnight S.L. Science. 1988; 240: 1759-1764Crossref PubMed Scopus (2544) Google Scholar). An amphipathic helix has a similar secondary structure but bears a hydrophobic face rather than a hydrophobic line. It appears that an advantage of the amphipathic helix is that it provides the opportunity for a variety of protein interactions to occur with different affinities and specificities. The isoform-specific binding of AKAPs leads us to hypothesize that the selection of a binding partner to an amphipathic helix is due to modulation of the binding surface by a change in the combination of hydrophobic amino acids with differently sized residues. Although more than 20 AKAPs have been reported in different tissues and species, there are still many questions about their roles in cells. In most cases the target proteins of the PKAs anchored by AKAPs are unknown, other proteins associated with AKAPs have not been identified, isoform-specific functions for RI and RII have not been determined, and the reasons for specific binding of RI or RII to AKAPs remain to be determined. The information gained in the current study should be useful for addressing some of these questions. The three amino acid rule forms the basis for testing the role of anchoring domains in other AKAPs. It will be informative to determine if the introduction of point mutations to abolish binding or to modify binding specificity of particular AKAPs in vivo causes physiological changes or different response to hormones and growth factors. This approach in conjunction with NMR and x-ray crystallography can be expected to give insight into the nature and role of interactions between AKAPs and PKA regulatory subunits at specific times and locations within cells and organisms. We thank Dr. Daniel W. Carr for providing the Ht31 cDNA. We also thank Dr. Yoshimitsu Kakuta and Deborah R. Blizard for helpful discussions.
Referência(s)