Transactivation of a human cytomegalovirus early promoter by gene products from the immediate-early gene IE2 and augmentation by IE1: mutational analysis of the viral proteins.

Expression from a human cytomegalovirus early promoter (E1.7) has been shown to be activated in trans by the IE2 gene products (C.-P. Chang, C. L. Malone, and M. F. Stinski, J. Virol. 63:281-290, 1989). Using wild-type and mutant viral proteins, we have defined the protein regions required for transactivation of the E1.7 promoter in IE2 and for augmentation of transactivation in the IE1 protein. Two regions of the IE2 proteins were found to be essential for transactivation. One near the amino terminus is within 52 amino acids encoded by exon 3. The second comprises the carboxyl-terminal 85 amino acids encoded by exon 5. The IE2 protein encoded by an mRNA which lacks the intron within exon 5 and the IE2 protein encoded by exon 5 had no activity for transactivation of the E1.7 promoter. Although the IE1 gene product alone had no effect on this early viral promoter, maximal early promoter activity was detected when both IE1 and IE2 gene products were present. The IE1 protein positively regulated its enhancer-containing promoter-regulatory region. The IE1 protein alone increased the steady-state level of IE2 mRNA; therefore, IE1 and IE2 are synergistic for expression from the E1.7 promoter. Like the IE2 proteins, the IE1 protein requires for activity 52 amino acids encoded by exon 3. IE1 also requires amino acids encoded by exon 4. Since the IE1 and IE2 proteins have 85 amino acids in common at the amino-terminal end encoded by exons 2 and 3, the difference between these specific transactivators resides in their carboxyl-terminal amino acids encoded by exons 4 and 5, respectively.     

GABAA receptors display association of gamma 2-subunit with alpha 1- and beta 2/3-subunits.

Recombinant GABAA (gamma-aminobutyrate-Type A) receptors that are sensitive to benzodiazepine receptor ligands can be generated by coexpression of alpha-, beta-, and gamma 2-subunit cDNAs (Pritchett, D. B., Sontheimer, H., Shivers, B. D., Ymer S., Kettenmann, H., Schofield, P. R., and Seeburg, P. H. (1989) Nature 338, 582-585; Pritchett, D. B., Lüddens, H., and Seeburg, P. H. (1989) Science 245, 1389-1392; Malherbe, P., Sigel, E., Baur, R., Perssohn, E., Richards, J. G., and Mohler, H. (1990) J. Neurosci. 10, 2330-2337). However, in brain tissue, only alpha- and beta-subunit proteins have so far been detected. To identify the size and distribution of the gamma 2-subunit protein in brain tissue, polyclonal antibodies were prepared against two synthetic peptides corresponding to amino acids 1-15 and 336-350 of the cDNA-derived rat gamma 2-subunit sequence. On Western blots, both anti-gamma 2-subunit antisera selectively labeled a 43-kDa protein. gamma 2-Subunit immunoreactivity was detected immunohistochemically in various brain regions, e.g. in the olfactory bulb, cerebral cortex, islands of Calleja, hippocampus, substantia nigra, and cerebellum. Immunoprecipitation with both antisera identified the gamma 2-subunit immunoreactivity in 40 and 50% of the native GABAA receptors purified from bovine and rat brains, respectively. Monoclonal antibody bd24 selectively recognizes the alpha 1-subunit, whereas bd17 recognizes both the beta 2- and beta 3-subunits (Ewert, M., Shivers, B. D., Lüddens, H., Mohler, H., and Seeburg, P. H. (1990) J. Cell Biol. 110, 2043-2048). Since either of these monoclonal antibodies (bd17 and bd24) precipitated approximately 90% of the GABAA receptors, the gamma 2-subunit is frequently associated with the alpha 1-subunit and the beta 2- and/or beta 3-subunit in vivo.     

Characterization of interaction sites in the Saccharomyces cerevisiae ribosomal stalk components

The interactions among the yeast stalk components (P0, P1alpha, P1beta, P2alpha and P2beta) and with EF-2 have been explored using immunoprecipitation, affinity chromatography and the two-hybrid system. No stable association was detected between acidic proteins of the same type. In contrast, P1alpha and P1beta were found to interact with P2beta and P2alpha respectively. An interaction of P0 with P1 proteins, but not with P2 proteins, was also detected. This interaction is strongly increased with the P0 carboxyl end, which is able to form a pentameric complex with the four acidic proteins. The P1/P2 binding site has been located between residues 212 and 262 using different C-terminal P0 fragments. Immunoprecipitation shows the association of EF-2 with protein P0. However, the interaction is stronger with the P1/P2 proteins than with P0 in the two-hybrid assay. This interaction improves using the 100-amino-acid-long C-end of P0 and is even higher with the last 50 amino acids. The data indicate a specific association of P1alpha with P2beta and of P1beta with P2alpha rather than the dimerization of the acidic proteins found in prokaryotes. In addition, they suggest that stalk assembly begins by the interaction of the P1 proteins with P0. Moreover, as functional interactions of the complete P0 were found to increase using protein fragments, the data suggest that some active sites are exposed in the ribosome as a result of conformational changes that take place during stalk assembly and function.     

The alpha1-beta-subunit interaction that modulates calcium channel activity is reversible and requires a competent alpha-interaction domain

High voltage-gated calcium channels consist of a pore-forming subunit (alpha(1)) and three nonhomologous subunits (alpha(2)/delta, beta, and gamma). Although it is well established that the beta-subunit promotes traffic of channels to the plasma membrane and modifies their activity, the reversible nature of the interaction with the alpha(1)-subunit remains controversial. Here, we address this issue by examining the effect of purified beta(2a) protein on Ca(V)1.2 and Ca(V)2.3 channels expressed in Xenopus oocytes. The beta(2a)-subunit binds to the alpha(1)-interaction domain (AID) in vitro, and when injected into oocytes, it shifts the voltage dependence of activation and increases charge movement to ionic current coupling of Ca(V)1.2 channels. This increase depended on the integrity of AID but was not abolished by bafilomycin, demonstrating that the alpha(1)-beta interaction through the AID site can take place at the plasma membrane. Furthermore, injection of beta(2a) protein inhibited inactivation of Ca(V)2.3 channels and converted fast inactivating Ca(V)2.3/beta(1b) channels to slow inactivating channels. Inhibition of inactivation required larger concentration of beta(2a) in oocytes expressing Ca(V)2.3/beta(1b) channels than expressing Ca(V)2.3 alone but reached the same maximal level as expected for a competitive interaction through a single binding site. Together, our data show that the alpha(1)-beta interaction is reversible in intact cells and defines calcium channels beta-subunits as regulatory proteins rather than stoichiometric subunits.     

T-type and N-type calcium channels of Xenopus oocytes: evidence for specific interactions with beta subunits.

We used amplifying effects of calcium channel beta subunits to identify endogenous calcium channels in Xenopus oocytes. Expression of rat brain beta 4 increased macroscopic endogenous current magnitude with a small effect on kinetics. In contrast, expression of rat brain/cardiac beta 2 produced a much larger increase in current magnitude and dramatically slowed current decay. Low concentrations of omega-conotoxin GVIA irreversibly blocked currents in both uninjected and beta 2-injected oocytes. Single channel recordings revealed both T- and N-type calcium channels with conductances of 9 and 18 pS, respectively, in uninjected oocytes and in oocytes expressing either beta subunit. Expression of either beta subunit slowed average current decay of T-type single channels. Slowing of T-type current decay by expression of beta 2 was due to reopening of the channels. N-type single channel average current decay showed little change with expression of beta 4, whereas expression of beta 2 slowed average current decay.     

Suramin interacts with the calmodulin binding site on the ryanodine receptor,RYR1

Apocalmodulin and Ca(2+) calmodulin bind to overlapping sites on the ryanodine receptor skeletal form, RYR1, but have opposite functional effects on channel activity. Suramin, a polysulfonated napthylurea, displaces both forms of calmodulin, leading to an inhibition of activity at low Ca(2+) and an enhancement of activity at high Ca(2+). Calmodulin binding motifs on RYR1 are also able to directly interact with the carboxy-terminal tail of the transverse tubule dihydropyridine receptor (DHPR) (Sencer, S., Papineni, R. V., Halling, D. B., Pate, P., Krol, J., Zhang, J. Z., and Hamilton, S. L. (2001) J. Biol. Chem. 276, 38237-38241). Suramin binds directly to a peptide that corresponds to the calmodulin binding site of RYR1 (amino acids 3609-3643) and blocks the interaction of this peptide with both calmodulin and the carboxyl-terminal tail of the DHPR alpha(1)-subunit. Suramin, added to the internal solution of voltage-clamped skeletal myotubes, produces a concentration-dependent increase in the maximal magnitude of voltage-gated Ca(2+) transients without significantly altering L-channel Ca(2+) channel conducting activity. Together, these results suggest that an interaction between the carboxyl-terminal tail of the DHPR alpha(1)-subunit with the calmodulin binding region of RYR1 serves to limit sarcoplasmic reticulum Ca(2+) release during excitation-contraction coupling and that suramin-induced potentiation of voltage-gated Ca(2+) release involves a relief of this inhibitory interaction.     

Analysis of the minimal specificity of CED-3 using a yeast transcriptional reporter system

Determination of the substrate specificity of site-specific proteases helps define their physiological roles. We developed a yeast-based system for defining the minimal substrate specificity of site-specific proteases, within the context of a protein. Using this system, we characterized the P4-P1 substrate specificity of the nematode apoptotic caspase CED-3. Apart from an absolute requirement for aspartate at the P1 position, CED-3 is a relatively promiscuous caspase capable of cleaving substrates bearing many amino acids at P4-P2 sites.     

Peptide substrate profiling defines fibroblast activation protein as an endopeptidase of strict Gly(2)-Pro(1)-cleaving specificity

Fibroblast activation protein (FAP) is a serine protease of undefined endopeptidase specificity implicated in tumorigenesis. To characterize FAP's P(4)-P(2)(') specificity, we synthesized intramolecularly quenched fluorescent substrate sets based on the FAP cleavage site in alpha(2)-antiplasmin (TSGP-NQ). FAP required substrates with Pro at P(1) and Gly or d-amino acids at P(2) and preferred small, uncharged amino acids at P(3), but tolerated most amino acids at P(4), P(1)(') and P(2)('). These substrate preferences allowed design of peptidyl-chloromethyl ketones that inhibited FAP, but not the related protease, dipeptidyl peptidase-4. Thus, FAP is a narrow specificity endopeptidase and this can be exploited for inhibitor design.     

Phosphorylation sites of bovine brain myelin basic protein phosphorylated with Ca2+-calmodulin-dependent protein kinase from rat brain

The phosphorylation sites of myelin basic protein from bovine brain were determined after phosphorylation with Ca2+-calmodulin-dependent protein kinase. Four phosphorylated peptides were selectively and rapidly separated by reversed-phase high-performance liquid chromatography. Partial sequencing of the phosphorylated peptides by automated Edman degradation revealed that Ca2+-calmodulin-dependent protein kinase phosphorylated serine-16, serine-70, and threonine-95 specifically, as well as serine-115, which is located on the experimental allergic encephalitogenic determinant of the protein. Of the four amino acid sequences determined, two sequences surrounding phosphorylated amino acids, -Lys-Tyr-Leu-Ala-Ser(P)16-Ala- and -Arg-Phe-Ser(P)115-Trp-Gly-, have both sides of each phosphoserine residue occupied by hydrophobic amino acids, and a basic amino acid, arginine or lysine, is located at the position 2 or 4 residues amino-terminal to the phosphoserine residue. In contrast, the two other sequences surrounding phosphorylated amino acids, -Tyr-Gly-Ser(P)70-Leu-Pro-Glu-Lys- and -Ile-Val-Thr(P)95-Pro-Arg-, have a basic amino acid at the position 2 or 4 residues carboxyl-terminal to the phosphoamino acid residue.     

The heparin binding domain of vitronectin is the region that is required to enhance insulin-like growth factor-I signaling

We have shown that vitronectin (Vn) binding to a cysteine loop sequence within the extracellular domain of the beta3-subunit (amino acids 177-184) of alphaVbeta3 is required for the positive effects of Vn on IGF-I signaling. When Vn binding to this sequence is blocked, IGF-I signaling in smooth muscle cells is impaired. Because this binding site is distinct from the site on beta3 to which the Arg-Gly-Asp sequence of extracellular matrix ligands bind (amino acids 107-171), we hypothesized that the region of Vn that binds to the cysteine loop on beta3 is distinct from the region that contains the Arg-Gly-Asp sequence. The results presented in this study demonstrate that this heparin binding domain (HBD) is the region of Vn that binds to the cysteine loop region of beta3 and that this region is sufficient to mediate the positive effects of Vn on IGF-I signaling. We provide evidence that binding of the HBD of Vn to alphaVbeta3 has direct effects on the activation state of beta3 as measured by beta3 phosphorylation. The increase in beta3 phosphorylation associated with exposure of cells to this HBD is associated with enhanced phosphorylation of the adaptor protein Src homology 2 domain-containing transforming protein C and enhanced activation MAPK, a downstream mediator of IGF-I signaling. We conclude that the interaction of the HBD of Vn binding to the cysteine loop sequence of beta3 is necessary and sufficient for the positive effects of Vn on IGF-I-mediated effects in smooth muscle cells.     

Coding and noncoding variation of the human calcium-channel beta4-subunit gene CACNB4 in patients with idiopathic generalized epilepsy and episodic ataxia

Inactivation of the beta4 subunit of the calcium channel in the mouse neurological mutant lethargic results in a complex neurological disorder that includes absence epilepsy and ataxia. To determine the role of the calcium-channel beta4-subunit gene CACNB4 on chromosome 2q22-23 in related human disorders, we screened for mutations in small pedigrees with familial epilepsy and ataxia. The premature-termination mutation R482X was identified in a patient with juvenile myoclonic epilepsy. The R482X protein lacks the 38 C-terminal amino acids containing part of an interaction domain for the alpha1 subunit. The missense mutation C104F was identified both in a German family with generalized epilepsy and praxis-induced seizures and in a French Canadian family with episodic ataxia. These coding mutations were not detected in 255 unaffected control individuals (510 chromosomes), and they may be considered candidate disease mutations. The results of functional tests of the truncated protein R482X in Xenopus laevis oocytes demonstrated a small decrease in the fast time constant for inactivation of the cotransfected alpha1 subunit. Further studies will be required to evaluate the in vivo consequences of these mutations. We also describe eight noncoding single-nucleotide substitutions, two of which are present at polymorphic frequency, and a previously unrecognized first intron of CACNB4 that interrupts exon 1 at codon 21.     

Asymmetric interactions between the acidic P1 and P2 proteins in the Saccharomyces cerevisiae ribosomal stalk.

The Saccharomyces cerevisiae ribosomal stalk is made of five components, the 32-kDa P0 and four 12-kDa acidic proteins, P1alpha, P1beta, P2alpha, and P2beta. The P0 carboxyl-terminal domain is involved in the interaction with the acidic proteins and resembles their structure. Protein chimeras were constructed in which the last 112 amino acids of P0 were replaced by the sequence of each acidic protein, yielding four fusion proteins, P0-1alpha, P0-1beta, P0-2alpha, and P0-2beta. The chimeras were expressed in P0 conditional null mutant strains in which wild-type P0 is not present. In S. cerevisiae D4567, which is totally deprived of acidic proteins, the four fusion proteins can replace the wild-type P0 with little effect on cell growth. In other genetic backgrounds, the chimeras either reduce or increase cell growth because of their effect on the ribosomal stalk composition. An analysis of the stalk proteins showed that each P0 chimera is able to strongly interact with only one acidic protein. The following associations were found: P0-1alpha.P2beta, P0-1beta.P2alpha, P0-2alpha.P1beta, and P0-2beta.P1alpha. These results indicate that the four acidic proteins do not form dimers in the yeast ribosomal stalk but interact with each other forming two specific associations, P1alpha.P2beta and P1beta.P2alpha, which have different structural and functional roles.     

Generation of Two Forms of the γ-Aminobutyric AcidA Receptor γ-2-Subunit in Mice by Alternative Splicing

gamma-Aminobutyric acidA (GABAA) receptors are multisubunit ligand-gated ion channels which mediate neuronal inhibition by GABA and are composed of at least four subunit types (alpha, beta, gamma, and delta). The gamma 2-subunit appears to be essential for benzodiazepine modulation of GABAA receptor function. In cloning murine gamma 2-subunits, we isolated cDNAs encoding forms of the subunit that differ by the insertion of eight amino acids. LLRMFSFK, in the major intracellular loop between proposed transmembrane domains M3 and M4. The two forms of the gamma 2-subunit are generated by alternative splicing, as demonstrated by cloning and partial sequencing of the corresponding gene. The eight-amino-acid insertion encodes a potential consensus serine phosphorylation site for protein kinase C. These results suggest a novel mechanism for the regulation of the GABAA receptor by protein phosphorylation.     

Computational evidence for the ligand selectivity to the alpha4beta2 and alpha3beta4 nicotinic acetylcholine receptors

The homology models of the alpha4beta2 and alpha3beta4 nicotinic acetylcholine receptors (nAChRs) suggest that the two nAChR subtypes are different in their ligand-binding pockets due to the non-conserved residues in the beta-subunits. The docking of nicotine, epibatidine, A-84543, and the two analogs of A-84543 ligands 1 and 2 to the homology models of alpha4beta2 and alpha3beta4 is presented. It is found that the protonated amino groups of these ligands bind to the alpha-subunits, whereas the remaining parts of the ligands bind to the beta-subunits. The two non-conserved amino acids Lys77 and Phe117 in the beta2-subunit corresponding to Ile77 and Gln117 in the beta4-subunit are identified to be the key players determining the binding modes of the ligands. We demonstrate how the increase in the number of the atoms connecting the pyrrolidine and pyridine rings in A-84543, 1, and 2, and an introduction of the alkynyl substituent in the pyridine ring affect the binding and shift the selectivity of these ligands toward the beta2-containing receptors. Further improvement in affinity and selectivity in this and other series of the ligands may be achieved by designing molecules that would specifically target the non-conserved regions in nAChRs.     

Differential regulation of cellular adhesion and migration by recombinant laminin-5 forms with partial deletion or mutation within the G3 domain of alpha3 chain

The basement membrane protein laminin-5 promotes cell adhesion and migration. The carboxyl-terminal G3 domain in the alpha3 chain is essential for the unique activity of laminin-5. To investigate the function of the G3 domain, we prepared various recombinant laminin-5 forms with a partially deleted or mutated G3 domain. The deletion of the carboxyl-terminal 28 amino acids (region III) markedly decreased the cell adhesion activity with a slight loss of the cell motility activity toward BRL and EJ-1 cells. This change was attributed to the loss of Lys-Arg-Asp sequence. Further deletion of 83 amino acids (region II) led to almost complete loss of the cell motility activity. All charged amino acid residues tested in this region were not responsible for the activity loss. These results suggest that the G3 domain contains two distinct regions that differently regulate cell adhesion and migration. Analysis of laminin-5 receptors showed that integrins alpha3beta1, alpha6beta1, and alpha6beta4 had different but synergistic effects on cell adhesion and migration on laminin-5. However, the structural change of the G3 domain appeared not to change integrin specificity. The present study demonstrates that the G3 domain in laminin-5 plays a central role to produce different biological effects on cells.     

Synaptic localization and presynaptic function of calcium channel beta 4-subunits in cultured hippocampal neurons

Neurotransmitter release is triggered by the influx of Ca(2+) into the presynaptic terminal through voltage gated Ca(2+)-channels. The shape of the presynaptic Ca(2+) signal largely determines the amount of released quanta and thus the size of the synaptic response. Ca(2+)-channel function is modulated in particular by the auxiliary beta-subunits that interact intracellularly with the pore-forming alpha(1)-subunit. Using retrovirus-mediated gene transfer in cultured hippocampal neurons, we demonstrate that functional GFP-beta(4) constructs colocalize with the synaptic vesicle marker synaptobrevin II and endogenous P/Q-type channels, indicating that beta(4)-subunits are localized to synaptic sites. Costaining with the dendritic marker MAP2 revealed that the beta(4)-subunit is transported to dendrites as well as axons. The nonconserved amino- and carboxyl-termini of the beta(4)-subunit were found to target the protein to the synapse. Physiological measurements in autaptic hippocampal neurons infected with green fluorescent protein (GFP)-beta(4) revealed an increase in both excitatory post-synaptic current amplitude and paired pulse facilitation ratio, whereas the GFP-beta(4) mutant, GFP-beta(4)(Delta50-407), which demonstrated a cytosolic localization pattern, did not alter these synaptic properties. In summary, our data suggest a pre-synaptic function of the Ca(2+)-channel beta(4)-subunit in synaptic transmission.     

Ahnak1 interaction is affected by phosphorylation of Ser-296 on Cavβ₂

Ahnak1 has been implicated in protein kinase A (PKA)-mediated control of cardiac L-type Ca(2+) channels (Cav1.2) through its interaction with the Cavβ(2) regulatory channel subunit. Here we corroborate this functional linkage by immunocytochemistry on isolated cardiomyocytes showing co-localization of ahnak1 and Cavβ(2) in the T-tubule system. In previous studies Cavβ(2) attachment sites which impacted the channel's PKA regulation have been located to ahnak1's proximal C-terminus (ahnak1(4889-5535), ahnak1(5462-5535)). In this study, we mapped the ahnak1-interacting regions in Cavβ(2) and investigated whether Cavβ(2) phosphorylation affects its binding behavior. In vitro binding assays with Cavβ(2) truncation mutants and ahnak1(4889-5535) revealed that the core region of Cavβ(2) consisting of Src-homology 3 (SH3), HOOK, and guanylate kinase (GK) domains was important for ahnak1 interaction while the C- and N-terminal regions were dispensable. Furthermore, Ser-296 in the GK domain of Cavβ(2) was identified as novel PKA phosphorylation site by mass spectrometry. Surface plasmon resonance (SPR) binding analysis showed that Ser-296 phosphorylation did not affect the high affinity interaction (K(D)≈35 nM) between Cavβ(2) and the α(1C) I-II linker, but affected ahnak1 interaction in a complex manner. SPR experiments with ahnak1(5462-5535) revealed that PKA phosphorylation of Cavβ(2) significantly increased the binding affinity and, in parallel, it reduced the binding capacity. Intriguingly, the phosphorylation mimic substitution Glu-296 fully reproduced both effects, increased the affinity by ≈2.4-fold and reduced the capacity by ≈60%. Our results are indicative for the release of a population of low affinity interaction sites following Cavβ(2) phosphorylation on Ser-296. We propose that this phosphorylation event is one mechanism underlying ahnak1's modulator function on Cav1.2 channel activity.