Site-directed mutagenesis reveals putative regions of protein interaction within the transmembrane domains of connexins.

Through cysteine-scanning mutagenesis, the authors have compared sites within the transmembrane domains of two connexins, one from the alpha-class (Cx50) and one from the beta-class (Cx32), where amino acid substitution disrupts the function of gap junction channels. In Cx32, 11 sites resulted in no channel function, or an aberrant voltage gating phenotype referred to as "reverse gating," whereas in Cx50, 7 such sites were identified. In both connexins, the sites lie along specific faces of transmembrane helices, suggesting that these may be sites of transmembrane domain interactions. In Cx32, one broad face of the M1 transmembrane domain and a narrower, polar face of M3 were identified, including one site that was shown to come into close apposition with M4 in the closed state. In Cx50, the same face of M3 was identified, but sensitive sites in M1 differed from Cx32. Many fewer sites in M1 disrupted channel function in Cx50, and those that did were on a different helical face to the sensitive sites in Cx32. A more in depth study of two sites in M1 and M2 of Cx32 showed that side-chain length or branching are important for maintenance of normal channel behavior, consistent with this being a site of transmembrane domain interaction.

Aberrant gating, but a normal expression pattern, underlies the recessive phenotype of the deafness mutant Connexin26M34T.

Mutations in the gene GJB2, encoding the gap junction protein Connexin26 (Cx26), are the most prevalent cause of inherited hearing loss, and Cx26M34T was one of the first mutations linked to deafness (Kelsell et al., 1997; Nature 387, 80-83). We report the first characterization of the gating properties of M34T, which had previously been reported to be nonfunctional. Although homotypic mutant channels did not produce detectable currents, heterotypic pairings with wtCx26 confirmed that M34T formed intercellular channels, although the gating properties were altered. Cx26M34T displayed an inverted response to transjunctional voltage (Vj), mediating currents that activate in a time- and Vj-dependent manner. These characteristics suggest that the channel population is only partially open at rest, consistent with previous reports that dye transfer in M34T-expressing cells is reduced or abolished (e.g., Thonnissen et al., Human Genet. 111, 190-197). To investigate the controversial recessive/dominant behavior of this mutant, we coexpressed M34T with wtCx26 RNA at equimolar levels, mimicking the situation in heterozygotic individuals. Under these conditions, M34T did not significantly reduce Cx26/Cx26 coupling, or alter the electrophysiological properties of the wt channels, consistent with the recessive nature of the allele. Overexpression of the mutant did have some inhibitory effects on conductance, possibly explaining some of the previous reports in exogenous expression systems and some patients. Consistent with its electrophysiological behavior, we also show that M34T localizes to cell junctions in both transfected HeLa cells and patient-derived tissue.

Dissection of the molecular basis of pp60(v-src) induced gating of connexin 43 gap junction channels.

Suppression of gap-junctional communication by various protein kinases, growth factors, and oncogenes frequently correlates with enhanced mitogenesis. The oncogene v-src appears to cause acute closure of gap junction channels. Tyr265 in the COOH-terminal tail of connexin 43 (Cx43) has been implicated as a potential target of v-src, although v-src action has also been associated with changes in serine phosphorylation. We have investigated the mechanism of this acute regulation through mutagenesis of Cx43 expressed in Xenopus laevis oocyte pairs. Truncations of the COOH-terminal domain led to an almost complete loss of response of Cx43 to v-src, but this was restored by coexpression of the independent COOH-terminal polypeptide. This suggests a ball and chain gating mechanism, similar to the mechanism proposed for pH gating of Cx43, and K+ channel inactivation. Surprisingly, we found that v-src mediated gating of Cx43 did not require the tyrosine site, but did seem to depend on the presence of two potential SH3 binding domains and the mitogen-activated protein (MAP) kinase phosphorylation sites within them. Further point mutagenesis and pharmacological studies in normal rat kidney (NRK) cells implicated MAP kinase in the gating response to v-src, while the stable binding of v-src to Cx43 (in part mediated by SH3 domains) did not correlate with its ability to mediate channel closure. This suggests a common link between closure of gap junctions by v-src and other mitogens, such as EGF and lysophosphatidic acid (LPA).

Core domain mutation (S86Y) selectively inactivates polyubiquitin chain synthesis catalyzed by E2-25K.

The mammalian ubiquitin conjugating enzyme known as E2-25K catalyzes the synthesis of polyubiquitin chains linked exclusively through K48-G76 isopeptide bonds. The properties of truncated and chimeric forms of E2-25K suggest that the polyubiquitin chain synthesis activity of this E2 depends on specific interactions between its conserved 150-residue core domain and its unique 50-residue tail domain [Haldeman, M. T., Xia, G., Kasperek, E. M., and Pickart, C. M. (1997) Biochemistry 36, 10526-10537]. In the present study, we provide strong support for this model by showing that a point mutation in the core domain (S86Y) mimics the effect of deleting the entire tail domain: the ability to form an E2 approximately ubiquitin thiol ester is intact, while conjugation activity is severely inhibited (>/=100-fold reduction in kcat/Km). The properties of E2-25K enzymes carrying the S86Y mutation indicate that this mutation strengthens the interaction between the core and tail domains: both free and ubiquitin-bound forms of S86Y-25K are completely resistant to tryptic cleavage at K164 in the tail domain, whereas wild-type enzyme is rapidly cleaved at this site. Other properties of S86Y-26K suggest that the active site of this mutant enzyme is more occluded than the active site of the wild-type enzyme. (1) Free S86Y-25K is alkylated by iodoacetamide 2-fold more slowly than the wild-type enzyme. (2) In assays of E2 approximately ubiquitin thiol ester formation, S86Y-25K shows a 4-fold reduced affinity for E1. (3) The ubiquitin thiol ester adduct of S86Y-25K undergoes (uncatalyzed) reaction with dithiothreitol 3-fold more slowly than the wild-type thiol ester adduct. One model to accommodate these findings postulates that an enhanced interaction between the core and tail domains, induced by the S86Y mutation, causes a steric blockade at the active site which prevents access of the incoming ubiquitin acceptor to the thiol ester bond. Consistent with this model, the S86Y mutation inhibits ubiquitin transfer to macromolecular acceptors (ubiquitin and polylysine) more strongly than transfer to small-molecule acceptors (free lysine and short peptides). These results suggest that unique residues proximal to E2 active sites may influence specific function by mediating intramolecular interactions.

Structure and function of ubiquitin conjugating enzyme E2-25K: the tail is a core-dependent activity element.

Individual members of the conserved family of ubiquitin conjugating enzymes (E2s) mediate the ubiquitination and turnover of specific substrates of the ubiquitin-dependent degradation pathway. E2 proteins have a highly conserved core domain of approximately 150 amino acids which contains the active-site Cys. Certain E2s have unique terminal extensions, which are thought to contribute to selective E2 function by interacting either with substrates or with trans-acting factors such as ubiquitin-protein ligases (E3s). We used the mammalian ubiquitin conjugating enzyme E2-25K in a biochemical test of this hypothesis. The properties of two truncated derivatives show that the 47-residue tail of E2-25K is necessary for three of the enzyme's characteristic properties: high activity in the synthesis of unanchored K48-linked polyubiquitin chains; resistance of the active-site Cys residue to alkylation; and an unusual discrimination against noncognate (nonmammalian) ubiquitin activating (E1) enzymes. However, the tail is not sufficient to generate these properties, as shown by the characteristics of a chimeric enzyme in which the tail of E2-25K was fused to the core domain of yeast UBC4. These and other results indicate that the specific biochemical function of the tail is strongly dependent upon unique features of the E2-25K core domain. Thus, divergent regions within the conserved core domains of E2 proteins may be highly significant for function. Expression of truncated E2-25K as a glutathione S-transferase (GST) fusion protein resulted in the apparent recovery of E2-25K-specific properties, including activity in chain synthesis. However, the catalytic mechanism utilized by the truncated fusion protein proved to be distinct from the mechanism utilized by the wild-type enzyme. The unexpected properties of the fusion protein were due to GST-induced dimerization. These results indicate the potential for self-association to modulate the polyubiquitin chain synthesis activities of E2 proteins, and indicate that caution should be applied in interpreting the activities of GST fusion proteins.

Substrate properties of site-specific mutant ubiquitin protein (G76A) reveal unexpected mechanistic features of ubiquitin-activating enzyme (E1).

Ubiquitin-mediated proteolysis proceeds via the formation and degradation of ubiquitin-protein conjugates. Ubiquitin (Ub)-activating enzyme (E1) catalyzes the first, MgATP-dependent step in the conjugative reaction sequence. With wild type ubiquitin, the product of the E1 reaction is a ternary complex (E1-Ub-AMP-Ub) containing one thiol-linked ubiquitin (via the Ub COOH terminus, Gly-76) and one tightly bound ubiquitin adenylate. The thiol-linked ubiquitin is subsequently transferred to the thiol of a ubiquitin-conjugating enzyme (E2 protein); the latter adduct is the proximal donor of ubiquitin to the target protein. A mutant ubiquitin, bearing a Gly to Ala substitution at the COOH terminus (G76A-ubiquitin), was characterized as a substrate for E1. G76A-ubiquitin 1) supported PPi-ATP exchange poorly (500-fold decrease in kcat/K(m); 2) did not produce detectable AMP-Ub with native E1; 3) produced stoichiometric AMP-Ub with thiol-blocked E1; 4) gave a stoichiometric burst of ATP consumption (1 mol/mol E1) with either native or thiol-blocked E1; 5) supported E1-ubiquitin thiol ester formation with native E1; 6) supported several downstream reactions of the proteolytic pathway at approximately 20% of the rate of wild type ubiquitin. These results indicate that G76A-ubiquitin gives a binary E1 thiol ester complex with native E1, due to the failure of the E1-ubiquitin thiol ester to undergo another round of adenylate synthesis; thus AMP-Ub is detected only if adenylate to thiol transfer is prevented by alkylating E1. The inability of G76A-ubiquitin to support ternary complex formation has implications for E1 active site structure. In other experiments, occupancy of the nucleotide/adenylate site of E1, by either MgATP or AMP-Ub, was found to stimulate ubiquitin transthiolation between E1 and E2 proteins. The intermediacy of ubiquitin adenylate thus provides a previously unrecognized catalytic advantage in the E1 mechanism.

Inhibition of ubiquitin-protein ligase (E3) by mono- and bifunctional phenylarsenoxides. Evidence for essential vicinal thiols and a proximal nucleophile.

Trivalent arsenoxides bind to vicinal thiol groups of proteins. We showed previously that the simplest trivalent arsenoxide, inorganic arsenite, inhibits ubiquitin-dependent protein degradation in rabbit reticulocyte lysate (Klemperer, N.S., and Pickart, C.M. (1989) J. Biol. Chem. 264, 19245-19242). We now show that, relative to arsenite, phenylarsenoxides are 10-165-fold more potent inhibitors of protein degradation in the same system (K0.5 for inhibition by p-aminophenylarsenoxide was 3.5-20 microM, depending on the substrate). In the ubiquitin-dependent proteolytic pathway, covalent ligation of ubiquitin to protein substrates targets the latter for degradation. In certain cases, specificity in ubiquitin-substrate conjugation depends critically upon the properties of ubiquitin-protein ligase or E3. Among other effects, p-aminophenylarsenoxide decreased the steady-state level of ubiquitinated human alpha-lactalbumin; this is a substrate which is acted upon directly by ubiquitin-protein ligase-alpha (E3-alpha). This finding suggests that phenylarsenoxides (unlike arsenite) inhibit E3. Several other lines of evidence confirm this conclusion. 1) A complex of E3-alpha and the 14-kDa ubiquitin-conjugating (E2) isozyme binds to phenylarsenoxide-Sepharose resin, with the E3 component of the complex mediating binding. 2) p-Aminophenylarsenoxide inhibited isolated E3 (K0.5 approximately 50 microM); inhibition was readily reversed by addition of dithiothreitol (which contains a competing vicinal thiol group), but not by beta-mercaptoethylamine (a monothiol). 3) A bifunctional phenylarsenoxide (bromoacetylaminophenylarsenoxide) rapidly and irreversibly inactivated E3; bromoacetyl aniline, which lacks an arsenoxide moiety, did not inhibit E3. These results suggest that E3 possesses essential vicinal thiol groups and that there is a reactive nucleophile proximal to the vicinal thiol site. The bifunctional phenylarsenoxide should be a useful tool for probing the relationship between structure and function in E3. As expected from prior results with arsenite, p-aminophenylarsenoxide was also a potent inhibitor of the turnover of ubiquitin-(human) alpha-lactalbumin conjugates.

Iodination of tyrosine 59 of ubiquitin selectively blocks ubiquitin's acceptor activity in diubiquitin synthesis catalyzed by E2(25K).

Covalent ligation of multiubiquitin chains targets eukaryotic proteins for degradation. Ubiquitin-conjugating enzyme E2(25K) utilizes isolated ubiquitin as the substrate for synthesis of such chains, in which successive ubiquitin units are linked by isopeptide bonds involving the side chain of Lys-48 of one ubiquitin and the COOH group of Gly-76 of the next. During continuous synthesis of multiubiquitin chains in the presence of purified ubiquitin-activating enzyme and E2(25K), there was a slight discrimination against radioiodinated ubiquitin (2.3-fold reduction in specific radioactivity of diubiquitin relative to value expected for no discrimination). Single-turnover experiments employing stoichiometrically iodinated ubiquitin derivatives indicated that E2(25K) discriminates extremely strongly (greater than 20-fold reduction in kcat/Km for diubiquitin synthesis) against ubiquitin that is monoiodinated at Tyr-59. The modest overall selection effect observed in continuous reactions is in part due to the occurrence of discrimination only when iodotyrosylubiquitin is the acceptor (Lys-48 donor) in diubiquitin synthesis; iodotyrosylubiquitin is kinetically competent when it is the species being transferred to native ubiquitin. The competence as acceptor of a site-directed mutant form of ubiquitin bearing a Tyr to Phe substitution at position 59 indicated that discrimination against iodotyrosylubiquitin by E2(25K) is not due to loss of the hydrogen-bonding interactions of Tyr-59. Rather, iodotyrosylubiquitin may be unable to react with the ubiquitin adduct of E2(25K) for steric reasons. Discrimination against iodotyrosylubiquitin as acceptor is unique to E2(25K) among three enzymes surveyed: iodotyrosylubiquitin is a fully competent acceptor in diubiquitin synthesis catalyzed by E2(25K) and is also utilized for multiubiquitin chain synthesis by E2(14K) and ubiquitin-protein ligase. These findings should assist in the design of future studies concerning E2(25K) structure and function.