Molecular Characterization of an Anchor Protein (AKAPCE) That Binds the RI Subunit (RCE) of Type I Protein Kinase A from Caenorhabditis elegans*

Classical A kinase anchor proteins (AKAPs) preferentially tether type II protein kinase A (PKAII) isoforms to sites in the cytoskeleton and organelles. It is not known if distinct proteins selectively sequester regulatory (R) subunits of type I PKAs, thereby diversifying functions of these critical enzymes. InCaenorhabditis elegans, a single type I PKA mediates all aspects of cAMP signaling. We have discovered a cDNA that encodes a binding protein (AKAPCE) for the regulatory subunit (RCE) of C. elegans PKAICE. AKAPCE is a novel, highly acidic RING finger protein composed of 1,280 amino acids. It binds RI-like RCE with high affinity and neither RIIα nor RIIβ competitively inhibits formation of AKAPCE·RCE complexes. The RCE-binding site was mapped to a segment of 20 amino acids in an N-terminal region of AKAPCE. Several hydrophobic residues in the binding site align with essential Leu and Ile residues in the RII-selective tethering domain of prototypic mammalian AKAPs. However, the RCE-binding region in AKAPCEdiverges sharply from consensus RII-binding sites by inclusion of three aromatic amino acids, exclusion of a highly conserved Leu or Ile at position 8 and replacement of C-terminal hydrophobic amino acids with basic residues. AKAPCE·RCE complexes accumulate in intact cells.

Classical A kinase anchor proteins (AKAPs) preferentially tether type II protein kinase A (PKAII) isoforms to sites in the cytoskeleton and organelles. It is not known if distinct proteins selectively sequester regulatory (R) subunits of type I PKAs, thereby diversifying functions of these critical enzymes. In Caenorhabditis elegans, a single type I PKA mediates all aspects of cAMP signaling. We have discovered a cDNA that encodes a binding protein (AKAP CE ) for the regulatory subunit (R CE ) of C. elegans PKAI CE . AKAP CE is a novel, highly acidic RING finger protein composed of 1,280 amino acids. It binds RI-like R CE with high affinity and neither RII␣ nor RII␤ competitively inhibits formation of AKAP CE ⅐R CE complexes. The R CE -binding site was mapped to a segment of 20 amino acids in an N-terminal region of AKAP CE . Several hydrophobic residues in the binding site align with essential Leu and Ile residues in the RII-selective tethering domain of prototypic mammalian AKAPs. However, the R CE -binding region in AKAP CE diverges sharply from consensus RII-binding sites by inclusion of three aromatic amino acids, exclusion of a highly conserved Leu or Ile at position 8 and replacement of Cterminal hydrophobic amino acids with basic residues. AKAP CE ⅐R CE complexes accumulate in intact cells.
A kinase anchor proteins (AKAPs) 1 avidly bind regulatory subunits (RII␤ and RII␣) of the type II isoforms of cAMP-dependent protein kinase (PKAII␤ and PKAII␣) (1)(2)(3). The various AKAPs are functional and not structural homologs; they possess partly homologous RII-binding domains, but the remaining portions of their sequences are divergent. Moreover, different AKAPs target PKAII isoforms to distinct intracellular loci including mitochondria, peroxisomes, Golgi, and plasma membranes and sites in cytoskeleton (1)(2)(3)(4)(5)(6). This accomplishes incorporation of AKAP⅐PKAII complexes into signaling modules that are either juxtaposed with or spatially segregated from hormone-activated adenylate cyclase. Substrate-effector proteins are thought to be clustered in proximity with anchored PKAII.
RI␣ is expressed in most tissues, where it mediates (as PKAI␣) actions of many hormones. It is the only R isoform that is essential for the progression of embryogenesis (7). Persistent synthesis of low levels of free RI␣ serves as a critical homeostatic mechanism to maintain the total R:PKA catalytic subunit (C) ratio at 1 (7). If excess C accumulates it is bound by RI␣, and the t1 ⁄2 of RI␣ increases 5-fold, thereby inhibiting otherwise unregulated and potentially toxic C activity. It is possible that RI␣ (PKAI␣) functions are further diversified by interactions with anchor proteins. This topic received little attention because fractionation of cell and tissue homogenates and immunocytochemical localization analysis indicate that the bulk of PKAI is dispersed in cytoplasm. However, RI is bound to the plasma membrane of human erythrocytes, and PKAI is recruited to a multi-protein complex at the "cap" site of activated T lymphocytes (8,9). Proteins and mechanisms involved in sequestration of PKAI in blood cells have not been characterized. However, high affinity RII tethering sites in several AKAPs also sequester RI␣ subunits with low avidity (10 -12). Huang et al. (10) reported in vitro binding of RII␣ or RI␣ with a single site on an isoform of S-AKAP84 (4) known as D-AKAP1 (10) or AKAP121 (6). S-AKAP84 targets and tethers PKAII to the outer membrane of mitochondria and contains the same RII/RI-binding site as D-AKAP1/AKAP121 (4,6). Moreover, high concentrations of RI␣ competitively inhibit the binding of RII subunits with D-AKAP1 and Ht31 (10,11). Although RII␣ is bound with 25-500-fold higher affinity than RI␣ in the cited studies (10,11), the results suggest that (a) RI dimers have a site available for interactions with potential anchor proteins and (b) some PKAI molecules may be immobilized within cells.
The preceding observations raise additional pertinent questions. Are there high affinity, RI-specific anchor proteins? If so, what is the biochemical basis for the binding and anchoring of RI (PKAI)? To what degree are structures of RI anchor proteins homologous with or divergent from RII-selective AKAPs? The nematode Caenorhabditis elegans provides an advantageous system for addressing these questions. The pathways, protein components, and mechanisms of cell signaling systems are conserved between C. elegans and mammals (13,14). However, the C. elegans genome contains only one C gene and one R gene (15,16). C. elegans C is 82% identical with mammalian C␣. C. elegans R (R CE ) is closely related to mammalian RI␣ (ϳ60% overall identity) but not the RII isoforms (ϳ30% identity) (16). Like RI␣ (17), R CE has a pseudosubstrate site and two N-* This work was supported in part by National Institutes of Health Grant GM22792 (to C. S. R.) and a National Institutes of Health Pharmacological Sciences Training Grant GM07260 stipend (to R. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The terminal Cys residues that covalently cross-link R CE subunits in vitro. 2 Moreover, R CE has (a) an N-terminal region that is homologous with the RI dimerization domain and (b) cAMPbinding sites that are 100% (site 1) and 87% (site 2) identical with corresponding regions in RI␣ (16). These properties strongly indicate that R CE is a homolog of RI.
A basic conundrum is as follows: how can one PKA holoenzyme mediate all aspects of cAMP-regulated signaling? Fractionation of C. elegans homogenates revealed that only 40% of PKA is isolated in cytosol, whereas 60% partitioned with the particulate fraction (16). Therefore, R CE anchor proteins might play a key role in diversifying and adapting C. elegans PKA for multiple functions. We now report the discovery and characterization of an R CE -binding protein, AKAP CE .

EXPERIMENTAL PROCEDURES
Isolation of cDNAs Encoding AKAP CE -Approximately 500,000 plaques from a C. elegans cDNA library in the expression vector ZAPII (CLONTECH) were screened by the procedure of Bregman et al. (18,19). Full-length, 32 P-labeled R CE (1.5 ϫ 10 5 cpm/ml) (see below) was used as a probe. Complexes of radiolabeled R CE with ␤-galactosidase-AKAP fusion proteins were detected by autoradiography (18). Positive plaques were purified to homogeneity. Recombinant pBluescript SK phagemids that contained anchor protein cDNAs were excised from the phage in Escherichia coli (20), amplified, and purified. Two related cDNA clones were isolated. The larger cDNA (4.3 kbp) contained the complete coding and 3Ј-untranslated sequences for AKAP CE mRNA, as well as a segment of the 5Ј-untranslated region. The entire sequence of the smaller cDNA (3.3 kbp) is included within the larger insert.
DNA Sequence Analysis-cDNA inserts were sequenced by a dideoxynucleotide chain termination procedure (21) as described previously (6).
Computer Analysis-Analysis of sequence data, sequence comparisons, and data base searches were performed using PCGENE-Intelligenetics software (Intelligenetics, Mountainview, CA) and BLAST programs (22,23) provided by the NCBI server at the National Institutes of Health.
Expression, Purification, Phosphorylation, and Mutagenesis of C. elegans R CE -Complementary DNA encoding C. elegans R CE was cloned and sequenced as described in a previous paper (16). A DNA insert that encodes amino acids 1-73 in R CE was amplified by the Pfu DNA polymerase (Stratagene) chain reaction (PCR). The 5Ј primer contained a NotI restriction sequence upstream from an ATG codon in the context of an NdeI cleavage site. This was followed by 18 nucleotides corresponding to codons 2-7 of R CE (16). The 3Ј primer corresponded to the complement of codons 67-73 in R CE . A unique BspE1 site is embedded in codon 71. Amplified DNA was digested with NotI and BspE1 and cloned into a recombinant pBluescript plasmid (which contains full-length R CE cDNA) that was cleaved with NotI (a 5Ј vector site) and BspE1. Fidelity of the PCR amplification was verified by DNA sequencing. Next, fulllength R CE cDNA was released from pBluescript by digestion with NdeI and KpnI (a 3Ј vector site) and cloned into the plasmid pGEM3Z (Promega) that was digested with the same restriction enzymes. Finally, R CE cDNA was excised from the pGEM3Z plasmid by cleavage with NdeI and BamHI (a 3Ј vector site) and cloned into the expression plasmid pET14b (Novagen), which was cut with the same restriction endonucleases. In pET14b R CE cDNA (375 codons) is preceded by DNA that encodes a 20-residue N-terminal fusion peptide that includes 6 consecutive His residues. Transcription of the chimeric gene is governed by a promoter sequence for bacteriophage T7 RNA polymerase, and synthesis of the polymerase is induced with isopropyl-1-thio-␤-Dgalactopyranoside (IPTG). E. coli BL21 (DE3) was transformed with the expression plasmid, and synthesis of His-tagged R CE was induced with 1 mM IPTG for 2 h at 37°C. Bacteria were harvested, disrupted, and separated into soluble and particulate fractions as described previously (24). The R CE fusion protein was purified to near homogeneity by Ni 2ϩ -chelate chromatography as described by Li and Rubin (25).
To generate a radiolabeled probe for screening a cDNA expression library (see above), purified R CE was phosphorylated by incubation with Mg-[␥-32 P]ATP and casein kinase as described previously (25). R CE was labeled to a higher specific activity for overlay and equilibrium binding assays by using a different approach. Gly 94 in the pseudosubstrate site (RRTGI) was replaced with a phosphorylatable residue (Ser) via sitedirected mutagenesis, using previously described conditions (26). This mutation has no effect on the properties of R CE . However, it enables the efficient phosphorylation of R CE by incubation with Mg-[␥-32 P]ATP and the catalytic subunit of protein kinase A, as previously reported (18).
Deletion Mutagenesis of AKAP CE -Deletion mutagenesis was performed via PCR as described for AKAP75 and S-AKAP84 (4,6,26). Mutants were verified by DNA sequencing. Amplified cDNAs were cloned into the expression plasmid pGEX-KG (27) after digestion of insert and vector with EcoRI and HindIII. This placed cDNAs encoding partial AKAP CE polypeptides downstream from and in-frame with the 3Ј terminus of a glutathione S-transferase (GST) gene in the vector. Transcription of the GST fusion gene is driven by an inducible tac promoter. High levels of chimeric GST partial AKAP CE polypeptides were produced when E. coli DH5␣ was transformed with recombinant pGEX-KG plasmids and then induced with 0.5 mM IPTG as described previously (28). Amino acid substitutions were introduced into the R CE -binding domain of AKAP CE via site-directed mutagenesis, as described previously (26).
Growth of C. elegans and AV-12 Cells-C. elegans was grown at 20°C as described previously (29). A cell line (AV-12) derived from a hamster subcutaneous tumor was grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum.
Transfection of AV-12 Cells-Full-length R CE and AKAP CE cDNA inserts were excised from recombinant, pBluescript plasmids by digestion with NotI and ApaI and ligated into the pCIS2 mammalian expression vector (30), which was cleaved with the same enzymes. This placed the cDNAs downstream from a powerful cytomegalovirus promoter and upstream from a poly(A) addition signal. Transient transfections of AV-12 cells were performed by calcium phosphate precipitation (31). Cells were harvested 60 h after transfection.
Expression and Purification of a GST Fusion Protein That Contains the R CE -binding Domain-A fragment of AKAP CE cDNA (nucleotides 526 -1227, Fig. 2A) that encodes amino acids 176 -409 in the anchor protein was cloned into the expression plasmid pGEX-KG. The methodology is described above for "Deletion Mutagenesis of AKAP CE ." This enabled synthesis of a GST-AKAP CE fusion protein in E. coli DH5␣ that was transformed with recombinant plasmid and induced with IPTG (24,27). After induction, bacteria were disrupted in a French press, and the soluble GST-AKAP CE (176 -409) protein was purified to near homogeneity by affinity chromatography on GSH-Sepharose 4B beads (Amersham Pharmacia Biotech) as previously reported (24).
Production and Purification of IgGs Directed against AKAP CE and R CE -Samples of AKAP CE (176 -409) fusion protein were injected into rabbits (0.4 mg initial injection; 0.2 mg for each of four booster injections) at Covance Laboratories (Vienna, VA) at 3-week intervals. Serum was collected at 3-week intervals after the first injection. Purified GST partial AKAP CE antigen was coupled to CNBr-Sepharose 4B as described by Bregman et al. (19) to yield a final concentration of 1.7 mg of protein bound per ml of resin. Antibodies that bind GST were eliminated by passing serum (2 ml) over a column of GST-Sepharose 4B (24). Adsorbed serum was applied to a column (2 ml) of partial AKAP CE (176 -409)-Sepharose 4B. The resin was washed extensively, and anti-AKAP CE IgGs were then isolated by successive elutions at pH 2.5 and pH 11.8, as described previously (24). Sufficient 1 M Na 2 HPO 4 or 0.5 M Hepes was added to column fractions to adjust the pH to ϳ7.5. The IgG concentration was estimated from the absorbance at 280 nm. IgGs were dialyzed against 10 mM sodium phosphate, pH 7.5, containing 0.15 M NaCl and 50% (v/v) glycerol and were stored at Ϫ20°C.
Preparation of antibodies directed against R CE is described in Lu et al. (16). Anti-R CE IgGs were purified from serum by affinity chromatography on a column of Sepharose 4B that was derivatized with 2.2 mg/ml His-tagged, full-length R CE (see above). Specific anti-R CE IgGs were eluted and stored as described above.
Overlay Assay for R CE Binding Activity-Overlay binding assays have been described in several papers (18,19). In brief, a Western blot 2 R. Angelo and C. S. Rubin, unpublished results.
is probed with 32 P-labeled R CE (using a subunit concentration of 0.3-0.7 nM and 2 ϫ 10 5 cpm 32 P radioactivity/ml). Complexes of 32 P-R CE and R CE -binding proteins are visualized by autoradiography. Results were quantified by scanning laser densitometry (Amersham Pharmacia Biotech XL laser densitometer) or PhosphorImager analysis (Molecular Dynamics) as described previously (26,33).
Equilibrium Binding Assays-Aliquots (200 ng) of highly purified, GST partial AKAP CE protein (residues 176 -409, Fig. 2A) were used for binding assays. Residues 236 -255 constitute the high affinity R CEbinding site in AKAP CE (see "Results"). Full-length C. elegans R CE was expressed, purified, and labeled with 32 P as described above. Binding assays were performed in 250 l of buffer A (20 mM Tris-HCl, pH 7.4, 0.15 M NaCl, 0.5 mM EDTA, 1 mM dithiothreitol) supplemented with 5% (w/v) fat-free powdered milk proteins. The amount of 32 P-labeled R CE was varied from 0.1 to 25 nM. Incubations were carried out at 0°C for 90 min. Subsequently, 40 l of a 50% suspension of GSH-Sepharose 4B beads, which was pre-equilibrated with buffer A plus 5% milk proteins, was added and the incubation was continued for 30 min. Next, the samples were diluted to 1 ml with buffer A plus 5% milk, and the beads were pelleted by centrifugation at 2,000 ϫ g for 5 min at 4°C. The beads were then washed twice by resuspension in 1 ml of buffer A, 5% milk and centrifugation at 2,000 ϫ g at 4°C. Finally, partial AKAP CE ⅐R CE complexes on the beads were washed three times by resuspension in 1 ml of buffer A containing 0.1% Triton X-100 and centrifugation at 2,000 ϫ g. Bound 32 P-labeled R CE was eluted from the beads in 0.2 ml of 1% SDS, and radioactivity was determined by a scintillation counter. 32 P radioactivity in an aliquot of the first supernatant solution was also determined to measure the amount of free R CE . Similar results were obtained when the beads were collected and rapidly washed on glass fiber filters.
Immunoprecipitation of AKAP CE ⅐R CE Complexes-Cytosol was prepared from transfected AV-12 cells as described previously (31). Affinity purified IgGs directed against R CE (3 l, 1:1.5, relative to serum) or preimmune serum (2 l) were mixed with ϳ100 l of cytosol containing 150 g of protein, and the samples were incubated for 1 h at 4°C. Next, 30 l of a 50% (v/v) suspension of protein A-Sepharose 4B beads (Amersham Pharmacia Biotech) was added, and the samples were rotated at 4°C for 16 h. Subsequently, samples were diluted with 0.9 ml of buffer A containing 0.1% Tween 20 and centrifuged at 4,000 ϫ g for 5 min at 4°C. The beads were collected and washed five times (1 ml per wash) with buffer A, 0.1% Triton X-100 by resuspension and centrifugation at 4,000 ϫ g. Proteins in immune complexes bound to protein A-Sepharose 4B were dissolved and denatured in gel loading buffer, fractionated according to size by denaturing electrophoresis, and transferred to an Immobilon P membrane as described above. The Western blots were probed with anti-AKAP CE IgGs and AKAP CE ⅐IgG complexes were visualized as described above.

RESULTS
Isolation and Characterization of cDNA Encoding a Novel AKAP from C. elegans-The overlay binding assay described by Bregman et al. (18,19) detects ␤-galactosidase-AKAP fusion proteins in phage plaques by their ability to bind cognate, 32 P-labeled R subunits. To execute functional screening of a C. elegans cDNA expression library, it was essential to produce and purify the nematode RI-like, cAMP-binding subunit, R CE . A cDNA that encodes full-length R CE (375 amino acids) (16) was cloned into the expression plasmid pET14b. This enabled high level synthesis (in E. coli) of a soluble fusion protein in which the R CE polypeptide is preceded by a 20-residue Nterminal peptide encoded by plasmid DNA. The fusion peptide contains a block of six consecutive His residues that constitute a divalent metal-binding site. His 6 -R CE fusion protein was purified to near homogeneity by affinity chromatography on a Ni 2ϩ -chelate resin (Fig. 1). Approximately 1.5 mg of purified fusion protein was isolated from a 250-ml culture of E. coli.
Purified R CE was phosphorylated by incubation with Mg-[␥-32 P]ATP and casein kinase, and a low concentration of radiolabeled R CE dimers (ϳ10 Ϫ9 M) was used to screen a C. elegans cDNA expression library in bacteriophage ZAPII. Two recombinant phage clones that encode proteins which sequester 32 Plabeled R CE were retrieved from the library and purified to homogeneity. Plasmids (pBluescript SK) containing the 3.3-and 4.3-kbp cDNA inserts were obtained by helper phagemediated excision in E. coli (20). DNA sequence analysis revealed that the complete sequence of the smaller cDNA was included in the 4.3-kbp insert.
The 4,294-bp sequence of the larger cloned cDNA (named AKAP CE cDNA) is presented in Fig. 2A. The DNA sequence has a long open reading frame that begins with a Met codon (nucleotides 1-3) in a C. elegans consensus context for translation initiation ((A/G)NNATGG) and ends with a translation termination codon at nucleotides 3841-3843. A 3Ј-untranslated region composed of 418 nucleotides and a polyadenylate tail follow the translation stop codon ( Fig. 2A). Processing of the 3Ј end of AKAP CE mRNA is evidently controlled by either of two nested copies of a variant poly(A) addition signal (GAAGAA, nucleotides 4236 -4244) that precede the polyadenylate tail by 15 or 18 nucleotides.
Since only 16 nucleotides in the 5Ј-untranslated region of AKAP CE cDNA were determined by sequencing, it was necessary to verify assignment of the initiator ATG codon by further analysis. This task was facilitated by the recent determination of the sequence of a 41-kbp genomic DNA insert from cosmid D1022 (GenBank accession number U23517) by the C. elegans genome project consortium (34). The central portion of this genomic DNA fragment includes the entire AKAP CE gene (named kap-1) 3 and long segments of contiguous 5Ј-and 3Јflanking DNA. In the kap-1 gene, a consensus acceptor site (TTCTTGCAG; nucleotides 12-20, Fig. 2B) for the trans-splicing machinery of C. elegans is evident 5 bp upstream from the putative initiator ATG codon. The 5Ј-untranslated regions of many C. elegans mRNAs are covalently modified by incorporation of a 22-nucleotide, non-translated "leader" sequence that is donated from a distinct, 100-nucleotide spliced leader 1 RNA transcript (35). The leader RNA fragment is inserted immediately downstream from the acceptor site in a reaction that also causes excision of the original 5Ј end of the transcript, shortening of the 5Ј-untranslated region (to 27 nucleotides in the case of AKAP CE mRNA), and introduction of a trimethylguanosine cap at the 5Ј terminus of the mRNA (35,36). These modifications enhance gene expression by increasing the efficiency of translation and/or stability of the mRNA (35, 36). The identity of the 54-kDa R CE polypeptide was verified by documenting its ability to bind [ 3 H]cAMP and anti-R CE IgGs, as described previously (16). The gel was stained with Coomassie Blue.
Seven of 11 nucleotides at the extreme 5Ј terminus of the experimentally determined AKAP CE cDNA sequence are not encoded by the kap-1 gene (Fig. 2B). However, the first 11 nucleotides in the cDNA are (a) identical with nucleotides Structural Properties of AKAP CE -The derived amino acid sequence of AKAP CE is presented in Fig. 2A. AKAP CE is a novel, highly acidic (pI ϳ4.4) protein composed of 1,280 amino acids (M r ϭ 144,000). There are no homologs of AKAP CE in current data bases. However, several regions of potential functional significance are evident in the AKAP CE sequence. Residues 544 -596 constitute a RING finger motif (37) (CX 2 CX 11 CX 2 CX 4 CX 2 CX 23 CX 2 C, where C indicates Cys and X indicates any amino acid; the lengths of sequences between C2 and C3, and C6 and C7 are long and variable, typically 9 -39 residues and 4 -48 residues, respectively). RING finger domains fold into a configuration in which zinc is tightly bound, via tetrahedral coordination with 4 Cys residues (37). Thus, each RING domain sequesters two zinc atoms. The RING finger region of AKAP CE is flanked by two candidate PKA phosphorylation sites (residues 522-526 and 613-617, respectively) ( Fig. 2A). A third PKA target site is located between residues 711 and 716. A partial protein containing residues 445-745 of AKAP CE is a substrate for PKA in vitro. 2 Finally, a binding site for R CE was identified and characterized as described below.
Organization of the AKAP CE (kap-1) Gene-During the course of our investigations, the C. elegans genome project (34) deposited DNA sequences for large segments of chromosome II in the GenBank TM data base. A data base search with 5Ј and 3Ј AKAP CE cDNA sequences revealed that cosmid D1022 contained the complete AKAP CE (kap-1) gene. (No previous studies have addressed experimentally any aspects of kap-1 gene expression or properties and functions of the AKAP CE protein.) Alignment of the complete sequence of AKAP CE cDNA with the DNA sequence of cosmid D1022 disclosed the intron/exon organization of the cognate gene (Table I). The remarkably compact C. elegans AKAP CE structural gene contains 17 exons but spans only 5.3 kbp of DNA because 15 of the 16 introns are Յ58 bp in length (Table I). The AKAP CE locus lies slightly to the right of center on chromosome II, where it is flanked by genes named col-6 and sro-1 (34).
Residues 236 -255 in C. elegans AKAP CE Constitute an R CEbinding Site-In order to verify the R CE binding activity of AKAP CE and initiate determination of structural features that constitute the tethering site, several fragments of AKAP CE (segments composed of residues 176 -475, 445-745, 711-1011, and 976 -1280, Figs. 2A and 3A) were expressed as GST fusion proteins. Partial AKAP CE (176 -475) avidly bound 32 P-R CE , whereas the other anchor protein fragments had no tethering activity (Fig. 3B). Subsequently, N-and C-terminal deletion mutagenesis was employed to establish boundaries for the R CEbinding domain according to the strategy outlined in Fig. 3A. All GST fusion proteins that bind R CE contain amino acids 236 -255 (Fig. 3, B-D). This region has an overall organization that is somewhat similar to the RII-binding site in AKAP75, although little sequence identity is evident. Four aliphatic hydrophobic amino acids, including three that are essential for RII binding in AKAP75 (26), align with hydrophobic residues in the AKAP CE -binding domain (Fig. 4A). The putative tethering domain of AKAP CE is also predicted to be an amphipathic ␣-helix with a hydrophobic face (Fig. 4B). The AKAP CE tethering domain includes a Tyr and 2 Phe residues, which may be involved in recognition and binding of RI-like proteins. Aromatic amino acids are absent from RII-binding sites of most previously characterized AKAPs. Moreover, the rare Phe or Tyr residues (no more than one per binding site) observed in RII tethering regions have not been implicated in ligand binding activity.
Functional roles for selected, individual amino acids were evaluated by site-directed mutagenesis (Fig. 5). Substitution of Leu 246 -Val 247 with Ala-Ala had little effect on R CE binding activity, whereas replacement of Val 247 -Ile 248 with Ala residues extinguished tethering activity. Thus, the conserved Ile at position 248 is essential for binding C. elegans PKA. Substitution of Phe 243 with Ala sharply reduced (but did not eliminate) R CE ⅐AKAP CE complex formation, suggesting that an aromatic hydrophobic side chain may be essential for the maintenance of high affinity for the R CE ligand. In contrast, replacement of conserved Ala 240 and Leu 252 with Ser and Ala, respectively, had no effect on the R CE binding activity of AKAP CE .
AKAP CE Binds R CE with High Affinity-If AKAP CE subserves targeting and tethering of C. elegans PKA in vivo, then the AKAP CE ⅐R CE complex should have a low dissociation constant, indicative of highly stable protein-protein interactions. This parameter was measured in solution by incubating GST partial AKAP CE (residues 176 -409) to equilibrium with various concentrations of 32 P-labeled R CE . Subsequently, GSH-Sepharose 4B beads were added and GST-AKAP CE ⅐ 32 P-R CE complexes were isolated and washed. 32 P radioactivity on the beads (bound R CE ) and in the supernatant solution (free R CE ) was determined by scintillation counting, and data were analyzed by the method of Scatchard (38). Representative results shown in Fig. 6A yield a K D value of 7 nM. Thus, R CE is tightly bound by AKAP CE . In typical eukaryotic cells total R subunit concentration is ϳ300 nM (39). If the concentration of AKAP CE is assumed to be 10 -100 nM (e.g. in a specific cell type), then Ͼ85% of the binding sites would be occupied with R CE (C. elegans PKA). (See Ref. 40 for discussion of such calculations.) Mammalian RII␣ and RII␤ do not competitively inhibit binding of R CE with AKAP CE (Fig. 6B).
Production and Purification of Antibodies Directed against AKAP CE -cDNA encoding residues 176 -409 of AKAP CE was cloned downstream from the GST gene in the expression plasmid pGEX-KG. After transformation of E. coli DH5␣ and induction with 0.5 mM IPTG, the soluble GST-AKAP CE -(176 -409) fusion protein was purified to homogeneity via affinity chromatography on GSH-Sepharose 4B (Fig. 7A). Approximately 4 mg of fusion protein was purified from a 1-liter culture of E. coli. Antibodies directed against the partial AKAP CE protein were produced in rabbits. IgGs in the antiserum yield a robust signal with 25 ng of AKAP CE partial protein (cleaved from the fusion protein by thrombin to eliminate GST epitopes) on a Western blot (Fig. 7B). An excess of structurally unrelated protein (R CE ) is not bound by the antibodies. Excess antigen eliminates the positive signals, and preimmune serum did not complex any polypeptides on the blot. IgGs that bind the C. elegans anchor protein were purified from serum by affinity chromatography on Sepharose 4B that was derivatized with partial AKAP CE protein (see "Experimental Procedures"). Affinity purified anti-AKAP CE IgGs bound a protein with an apparent M r of 200,000 in homogenates of C. elegans (Fig. 7C, lane 1). The antibodies also complexed a protein of the same size in extracts of hamster AV-12 cells that were transfected with an AKAP CE transgene (Fig. 7C, lane 3). No antigen was observed among proteins isolated from nontransfected AV-12 cells (Fig. 7C, lane 2). The 200-kDa antigen was not detected when Western blots of proteins from C. elegans and transfected AV-12 cells were probed with preimmune IgGs or anti-AKAP CE IgGs that were preincubated with excess partial AKAP CE (Fig. 7C, lanes 4 -7). The apparent molecular weight of AKAP CE was confirmed by in vitro translation. AKAP CE mRNA (generated from the cDNA template shown in Fig. 2A) directed the synthesis of a 35 S-labeled polypeptide that exhibited a M r of 200,000 in a denaturing polyacrylamide gel (data not shown). The discrepancy between the apparent molecular weight (200,000) and calculated molecular weight (144,000) values for AKAP CE may be due to the highly acidic nature of the anchor protein. AKAP CE will bind SDS poorly and migrate abnormally slowly in denaturing polyacrylamide gels. Mammalian AKAPs exhibit similar physico-chemical properties (4,6,18,19).
AKAP CE ⅐R CE Complexes Are Isolated from Transfected Cells-Extracts were prepared from AV-12 cells that were cotransfected with both AKAP CE and R CE transgenes. Samples were incubated with affinity purified IgGs directed against R CE , and immune complexes were isolated on protein A-Sepharose 4B beads. Proteins in immunoprecipitates were fractionated according to size by denaturing electrophoresis and transferred to a polyvinylidene difluoride membrane. The Western blot was probed with anti-AKAP CE IgGs. Substantial amounts of AKAP CE were co-immunoprecipitated with R CE and anti-R CE IgGs (Fig. 8, lanes 1 and 2). Lane 3 (Fig. 8) shows that AKAP CE directly precipitated by anti-AKAP CE IgGs is indistinguishable from the anchor protein isolated in complexes with R CE (Fig. 8,  lanes 1 and 2). Results were confirmed by 32 P-labeled R CE overlay binding assays (data not shown). No signals were observed when immunoprecipitates from control (non-transfected) AV-12 cells were assayed by the same procedures (Fig.  8, lane 4). AKAP CE was not precipitated by preimmune serum (Fig. 8, lane 5) or anti-R CE immune IgGs preincubated with excess antigen (Fig. 8, lane 6). Thus, stable AKAP CE ⅐R CE complexes are generated in the physiological context of intact cells. DISCUSSION We have discovered and characterized cDNAs that encode a C. elegans A kinase anchor protein, AKAP CE . AKAP CE avidly binds (K D ϭ 7 nM) the RI␣-like regulatory subunit (R CE ) of the  Fig. 2A); numbers on the right identify the C-terminal residues in the partial AKAP CE proteins. Asterisks indicate AKAP CE fragments that bind 32 P-labeled R CE . E. coli DH5␣ was transformed with recombinant pGEX plasmids to enable the IPTG-inducible expression of the indicated partial proteins. After induction, samples of total E. coli proteins (30 g) were size-fractionated in a 0.1% SDS-10% polyacrylamide gel and transferred to an Immobilon P membrane. Polypeptides that sequestered 32 P-R CE were identified by performing overlay binding assays sole PKA that is expressed in the nematode (16). Since neither mammalian RII␣ and RII␤ nor Drosophila RII 2 efficiently compete for the R CE -binding site in the C. elegans anchor protein, it appears that AKAP CE is the first example of an RI-specific targeting/tethering protein.
AKAP CE is a novel protein composed of 1280 amino residues (M r ϭ 144,000). However, the nematode anchor protein shares two overall properties that are characteristic features of mammalian AKAPs; the polypeptide is highly acidic (pI ϳ4.4) and exhibits an erroneously high apparent molecular weight (200,000) in denaturing polyacrylamide gels. Moreover, several potentially important motifs are evident within a segment of AKAP CE that spans amino acids 200 -725 ( Fig. 2A). Residues 544 -596 are predicted to fold into a classical RING finger structure (see "Results"). Initially, it was thought that RING finger proteins were sequence-specific DNA-binding proteins. A substantial body of evidence indicates that this is not the case (37). RING finger domains are present in many cytoplasmic, cytoskeleton-associated and integral membrane proteins.

FIG. 4. Conservation and divergence in the structures of R CE and mammalian RII-binding sites in AKAPs.
A, the R CE tethering site of AKAP CE is aligned with RII-binding site of AKAP75 (26). Asterisks mark the positions of conserved hydrophobic residues; single dots indicate positions of major differences between the two binding sites. B shows a helical wheel depiction of the orientation of hydrophobic and hydrophilic residues in the R CE -and RII-binding sites. Residues that contribute to an extended, largely hydrophobic surface are underlined.

FIG. 5. Identification of individual amino acids in AKAP CE
that are essential for the high affinity binding of R CE . Selected amino acid residues in the R CE -binding site of AKAP CE were altered by site-directed mutagenesis as indicated under "Experimental Procedures." The wild type and mutated R CE -binding domains were expressed and assayed as described under "Experimental Procedures" and the legend for Fig. 3. However, proteins were purified by chromatography on GSH-Sepharose 4B to eliminate fragments. Equal amounts of protein (0.3 g) were applied to each lane. A representative autoradiogram obtained from an overlay binding assay is shown. The experiment was repeated three times, and similar results were obtained in each instance. FIG. 6. AKAP CE binds R CE avidly and specifically. A, equilibrium binding of 32 P-labeled R CE with the binding domain of AKAP CE was determined as described under "Experimental Procedures" and under "Results." The data are plotted according to the method of Scatchard, and the K D was determined from the slope. B shows the ability of various amounts of non-radioactive R CE (diamonds), RII␣ (squares), and RII␤ (triangles) to inhibit the binding of 32 P-labeled R CE with the AKAP CE -binding domain. In the absence of the competitor, the AKAP CE fusion protein bound 41,000 cpm.
Emerging concepts are that (a) proteins containing RING fingers are usually components of multi-protein complexes and (b) the RING domain plays a key role in the organization and assembly of such complexes (37). A pertinent example of a RING finger protein is STE5, the scaffolding protein of yeast that sequesters and co-localizes ␤␥ subunits of a G protein, MEKK, MEK, and mitogen-activated protein kinase. This multi-protein complex couples a pheromone-generated signal at the cell surface to gene transcription in the nucleus (37). The RING finger domain in AKAP CE is not essential for binding R CE (Fig. 3 and text of "Results"). Observations and concepts summarized above suggest the speculation that this region of the anchor protein mediates immobilization of R CE ⅐AKAP CE complexes at specific intracellular sites and/or clustering of the complexes with other signaling proteins.
Three consensus PKA phosphorylation sites (residues 522-526, 613-617, 711-716, Fig. 2A) are included in a fragment of AKAP CE (residues 445-745) that is phosphorylated by the catalytic subunit of PKA in vitro. 2 Once the phosphorylation sites are precisely mapped and mutated, it should be possible to determine if PKA-mediated phosphorylation reversibly regulates the ability of AKAP CE to bind with "docking" proteins (or R CE ), thereby modulating scaffolding or targeting functions.
The high affinity R CE -binding site was mapped to a short segment of AKAP CE that encompasses amino acids 236 -256. Comparison of the R CE -binding region with the prototypic RII binding module of bovine AKAP75 disclosed several similarities. Binding domains contain ϳ20 amino acids, and both are predicted to fold as ␣-helices that contain one large, predominantly hydrophobic face. Studies on mammalian AKAP75 and AKAP79 indicate that a large hydrophobic surface at the tethering site and a complementary apolar surface on RII␤ and RII␣ dimers are essential for formation of stable AKAP⅐PKAII complexes (25,26,41). Binding of AKAP CE with R CE may also be driven by hydrophobic interactions. Although sequence identity between the aligned binding domains is minimal, three aliphatic hydrophobic amino acids (Leu 236 , Ile 248 , and Leu 252 ) of AKAP CE match large aliphatic hydrophobic residues that are essential for the RII binding activity of AKAP75 (26). Mutation of Ile 248 to Ala eliminated the tethering activity of AKAP CE ; the corresponding mutation in AKAP75 yields the same result. Immune complexes were isolated on protein A-Sepharose 4B beads, washed extensively, and analyzed by Western immunoblot analysis as indicated under "Experimental Procedures." Lanes 1-3, 5, and 6 received immunoprecipitated proteins from cytosol of AV-12 cells that were co-transfected with both the AKAP CE and R CE transgenes; lane 4 contained proteins precipitated from non-transfected AV-12 cells. The blot was probed with affinity purified IgGs directed against AKAP CE (1:4000, relative to serum), and IgG-AKAP CE complexes were detected via peroxidase-coupled secondary antibodies and an enhanced chemiluminescence procedure as described in Fig. 7 and "Experimental Procedures." Lane 6 was incubated with IgGs that were saturated with excess, purified antigen. Only the relevant portion of the immunoblot is shown. The specificity of the anti-AKAP CE IgGs is documented in Fig. 7C. After incubation with peroxidase-coupled secondary antibodies, AKAP CE -IgG complexes were visualized by an enhanced chemiluminescence procedure (26,31). Signals were recorded on x-ray film. C, samples (30 g) of proteins from C. elegans (lanes 1, 4, and 6), AV-12 cells transfected with a full-length AKAP CE transgene (lanes 3, 5, and 7) and control AV-12 cells (lane 2) were subjected to Western immunoblot analysis using a 7.5% denaturing polyacrylamide gel, as described above. Lanes 1-3 were probed with anti-AKAP CE IgGs (1:4000); lanes 4 and 5 were incubated with preimmune serum (1:1000); lanes 6 and 7 were probed with anti-AKAP CE IgGs in the presence of 3 g of GST partial AKAP CE antigen. AKAP CE -IgG complexes were visualized as described above and in Ref. 31. Thus Ile 248 and its highly conserved counterparts in AKAP75 and other AKAPs (1-5) apparently play a central role in maintaining the structural integrity R subunit-binding sites.
Other features of the R CE -binding domain diverge sharply from properties conserved in RII-selective mammalian AKAPs. Substitution of Leu 252 with Ala had little effect on the R CE tethering capacity of AKAP CE , whereas the comparable replacement in AKAP75 compromises binding activity. Positions occupied by Phe 253 and Ser 254 in AKAP CE are almost invariably filled by Leu, Ile, or Val in mammalian AKAPs. Moreover, mutagenesis of Phe 253 (to Ala) revealed that this aromatic hydrophobic residue is essential for maintenance of high affinity for R CE . Two additional aromatic amino acids (Tyr 237 and Phe 239 ) are included in the R CE binding module, whereas aromatic residues are excluded from most RII-binding AKAPs. Finally, basic amino acids (Arg 255 and Lys 256 ) at the C terminus of the binding region in AKAP CE align with hydrophobic amino acids in AKAP75 (Fig. 4). Mutagenesis and binding analyses are needed to establish or exclude roles for Tyr 237 , Phe 239 , Arg 255 , Lys 256 , and Ser 254 in controlling the affinity and selectivity of the R CE -binding domain and to define further unique features of an RI-selective binding module. Similarities and differences between the R-binding sites in AKAP CE and S-AKAP84/D-AKAP1 (4, 6, 10) are essentially the same as those described above for AKAP CE and AKAP75.
The discovery of the R CE -binding domain in a C. elegans protein, the recent identification and characterization of an RII-selective binding site in a Drosophila AKAP (28), and conservation of structural features shared with mammalian RII tethering sites (see above) indicate the R binding modules and R subunits co-evolved in a broad range of eukaryotes. It is possible that current R binding modules are derived from a common ancestor domain that was incorporated into a variety of polypeptides through the process of exon shuffling. Conservation of R binding modules during evolution also provides further support for the idea that anchoring/targeting of PKA isoforms plays a pivotal role in diversifying and adapting A kinases for a variety of physiological functions. This concept is especially pertinent for C. elegans. This nematode expresses only one type of PKA molecule (R CE2 C CE2 ), yet the kinase is nearly equally distributed between cytosolic and particulate (i.e. membrane/cytoskeleton) fractions of tissues (16). Moreover, C. elegans PKA must mediate a myriad of conserved regulatory functions that are operative in all eukaryotes. The demonstration that AKAP CE avidly binds R CE in intact cells (Fig. 8) suggests that interactions of PKA with anchoring/ targeting proteins(s) provide a mechanism for diversification and specialization in PKA signaling pathways in the nematode.
The cDNA expression library (3 ϫ 10 6 independent clones) has not been exhaustively screened for AKAPs (which are typically encoded by low abundance mRNAs) because (a) only one cDNA in three has the appropriate reading frame for translation, and (b) partial cDNA clones may lack the 20-residue R CE -binding site. Thus, it is possible that additional AKAPs mediate targeting and tethering of C. elegans PKAI to organelles and cytoskeleton. An alternative targeting mechanism involves the possibility that distinct domains in the large (144 kDa) AKAP CE polypeptide bind with protein ligands that accumulate in certain organelles or portions of cytoskeleton. Such interactions would enable the distribution of PKAI⅐AKAP CE complexes to multiple intracellular sites. These possibilities are currently under investigation.
Immunoprecipitation experiments revealed that some AKAP CE is recovered in cytosol. This does not preclude a tight association of the anchor protein with cytoskeleton in situ. Many components of highly organized cytoskeletal structures are isolated in cytosol (e.g. microtubule associated protein-2, paxillin, actin-binding proteins, 15-30% of AKAP75) when cells/tissues are disrupted in standard, low ionic strength buffer (as used in our studies) in the absence of certain stabilizing agents. The development of stabilizing conditions for C. elegans cytoskeleton and the application of confocal immunofluorescence microscopy should facilitate assessment of the distribution of AKAP CE among various subcellular fractions in future experimentation. It is also possible that AKAP CE is a "cytoplasmic anchor" or "scaffold" that either (a) directs cAMP signaling to substrate/effector proteins in cytoplasm, (b) inhibits PKA-mediated signaling at certain organelles or sites in cytoskeleton, or (c) co-assembles tethered PKA with other substrate and effector proteins that are directly bound to other domains of the anchor protein. The cloning of AKAP CE cDNA and the availability of the kap-1 gene and its flanking (regulatory) regions will now permit enhancement, inhibition, mislocalization, and ablation of AKAP CE in situ in C. elegans via recently developed molecular genetics technology. Such approaches may ultimately reveal the exact physiological roles for anchored PKA.