In vivo coassembly of a divergent beta-tubulin subunit (c beta 6) into microtubules of different function

alpha- and beta-Tubulin are encoded in vertebrate genomes by a family of approximately 6-7 functional genes whose polypeptide products differ in amino acid sequence. In the chicken, one beta-tubulin isotype (c beta 6) has previously been found to be expressed only in thrombocytes and erythroid cells, where it is assembled into a circumferential ring of marginal band microtubules. In light of its unique in vivo utilization and its divergent assembly properties in vitro, we used DNA transfection to test whether this isotype could be assembled in vivo into microtubules of divergent functions. Using an antibody specific to c beta 6, we have found that upon transfection this polypeptide is freely coassembled into an extensive array of interphase cytoplasmic microtubules and into astral and pole-to-chromosome or pole-to-pole microtubules during mitosis. Further, examination of developing chicken erythrocytes reveals that both beta-tubulins that are expressed in these cells (c beta 6 and c beta 3) are found as co-polymers of the two isoforms. These results, in conjunction with efforts that have localized various other beta-tubulin isotypes, demonstrate that to the resolution limit afforded by light microscopy in vivo microtubules in vertebrates are random copolymers of available isotypes. Although these findings are consistent with functional interchangeability of beta-tubulin isotypes, we have also found that in vivo microtubules enriched in c beta 3 polypeptides are more sensitive to cold depolymerization than those enriched in c beta 6. This differential quantitative utilization of the two endogenous isotypes documents that some in vivo functional differences between isotypes do exist.     

Sequencing of peptides and proteins from the carboxy terminus.

A new chemical method for carboxy-terminal (C-terminal) protein sequencing has been developed. This approach has been successfully used to sequence 5 residues of standard proteins and 5 to 10 residues of synthetic peptides at low nanomole levels. The sequencing procedure consists of converting the C-terminal amino acid into a thiohydantoin (TH) derivative, followed by transformation of the TH into a good leaving group by alkylation. Next, the alkylated TH is cleaved mildly and efficiently with (N = C V S)- anion, which simultaneously forms a TH on the newly truncated protein or peptide. Thus, after the initial TH derivatization, there is no return to a free carboxyl group at the C-terminus. An additional benefit of this method is that the alkylating moiety can be chosen with a variety of properties allowing for variation in the detection method. This chemistry has been adapted to automated protein sequencers with a cycle time of about 1 h.     

Temperature-sensitive mutations at the carboxy terminus of the alpha subunit of the Escherichia coli F1F0 ATP synthase.

Mutations were constructed in the a subunit of the F1F0 ATP synthase from Escherichia coli. Truncated forms of this subunit showed a temperature sensitivity phenotype. We conclude that the carboxy terminus of the a subunit is not involved directly with proton translocation but that it has an important structural role.     

Removal of MAP4 from microtubules in vivo produces no observable phenotype at the cellular level

Microtubule-associated protein 4 (MAP4) promotes MT assembly in vitro and is localized along MTs in vivo. These results and the fact that MAP4 is the major MAP in nonneuronal cells suggest that MAP4's normal functions may include the stabilization of MTs in situ. To understand MAP4 function in vivo, we produced a blocking antibody (Ab) to prevent MAP4 binding to MTs. The COOH-terminal MT binding domain of MAP4 was expressed in Escherichia coli as a glutathione transferase fusion protein and was injected into rabbits to produce an antiserum that was then affinity purified and shown to be monospecific for MAP4. This Ab blocked > 95% of MAP4 binding to MTs in an in vitro assay. Microinjection of the affinity purified Ab into human fibroblasts and monkey epithelial cells abolished MAP4 binding to MTs as assayed with a rat polyclonal antibody against the NH2-terminal projection domain of MAP4. The removal of MAP4 from MTs was accompanied by its sequestration into visible MAP4-Ab immunocomplexes. However, the MT network appeared normal. Tubulin photoactivation and nocodazole sensitivity assays indicated that MT dynamics were not altered detectably by the removal of MAP4 from the MTs. Cells progressed to mitosis with morphologically normal spindles in the absence of MAP4 binding to MTs. Depleting MAP4 from MTs also did not affect the state of posttranslational modifications of tubulin subunits. Further, no perturbations of MT- dependent organelle distribution were detected. We conclude that the association of MAP4 with MTs is not essential for MT assembly or for the MT-based functions in cultured cells that we could assay. A significant role for MAP4 is not excluded by these results, however, as MAP4 may be a component of a functionally redundant system.     

Molecular cloning of the carboxy terminus of a canine tracheobronchial mucin.

A cDNA library constructed from canine tracheal mRNA was screened with polyclonal antiserum specific to canine tracheal apomucin (CTM-A). Eight antibody reactive clones were isolated and purified to clonality. One of the clones, designated pCTM-A, had a 1.7 kb insert and included a single open reading frame with a poly (A)+ tail. The amino acid composition of the encoded protein was consistent with that expected for CTM-A. The fusion protein produced by cloning the 1.7 kb insert in the pMALc expression vector reacted with the purified anti-apomucin CTM-A antibody. Also, polyclonal antibodies raised to the purified protein product encoded by pCTM-A reacted with deglycosylated CTM-A confirming that this clone does indeed code for apomucin CTM-A. This is the first report of a cDNA encoding the C-terminus of a canine tracheal mucin.     

Fluorescence studies on a membrane-embedded peptide from the carboxy terminus of lipophilin

Fluorescence of an intramembranous polypeptide (T-3) derived from the carboxy-terminal sequence of lipophilin was studied in aqueous solution, detergent micelles, and lipid vesicles. In all cases, the fluorescence of the only Trp (211) was indicative of a hydrophobic, buried residue. Addition of lysophosphatidylcholine (LPC) or phosphatidylcholine (PC) gave Trp-211 a more hydrophobic, less quenching environment as compared to that in aqueous solution. Energy transfer between Trp and Tyr observed in aqueous solution was decreased by the addition of lipid or detergent. There was limited quenching by acrylamide both in the aqueous and in the lipid or detergent environments. However, PC or LPC further decreased this quenching. Cs+ and I- were even less accessible than acrylamide to Trp, further proving that the Trp was located inside the lipid bilayer. The quenching indicated that I- binds to positive charges of the protein located on the surface of the membrane. This, combined with knowledge of the sequence of lipophilin, suggested that Trp-211 was located within the membrane but was close to amino acid residues that are external to the bilayer.     

Identification of a peroxisomal targeting signal at the carboxy terminus of firefly luciferase

Translocation of proteins across membranes of the endoplasmic reticulum, mitochondrion, and chloroplast has been shown to be mediated by targeting signals present in the transported proteins. To test whether the transport of proteins into peroxisomes is also mediated by a peptide targeting signal, we have studied the firefly luciferase gene that encodes a protein transported to peroxisomes in both insect and mammalian cells. We have identified two regions of luciferase which are necessary for transport of this protein into peroxisomes. We demonstrate that one of these, region II, represents a peroxisomal targeting signal because it is both necessary and sufficient for directing cytosolic proteins to peroxisomes. The signal is no more than twelve amino acids long and is located at the extreme carboxy-terminus of luciferase. The location of the targeting signal for translocation across the peroxisomal membrane therefore differs from the predominantly amino-terminal location of signals responsible for transport across the membranes of the endoplasmic reticulum, chloroplast, or mitochondrion.     

Mutations affecting assembly of beta-tubulin localize to a region near the carboxyl terminus

The generation of Chinese hamster ovary cell lines that express assembly defective forms of beta-tubulin were isolated using selections based on reversion of conditional lethal or drug resistance phenotypes. Two such cell lines, D2 and 6H3, were chosen for further characterization because they contain beta-tubulin polypeptides that exhibit decreases in apparent molecular weight on two-dimensional gel electrophoresis. Analysis of the nucleic acid from these cell lines using both Southern and Northern procedures suggests a deletion in one of the beta-tubulin genes in each cell line. Localization of the missing sequence in D2 was first determined by tryptic peptide mapping by high performance liquid chromatography. Subsequently, the assignment was confirmed by constructing appropriate subclones of a wild type Chinese hamster ovary beta-tubulin cDNA for Southern analysis to demonstrate a failure to recognize characteristic hybridization patterns of the mutant tubulin gene. In the other revertant, 6H3, the deletion was detected on a Northern blot by differential hybridization of a 3' fragment of the cDNA to the beta-tubulin messages. The results indicate that D2 has an internal deletion whose approximate limits extend from amino acid residues 250 through 345. Cell line 6H3 has a deletion that begins near amino acid residue 330 and extends into the 3'-untranslated region of the gene.     

Synthesis of a human leukocyte interferon with a modified carboxy terminus in Escherichia coli

A recombinant plasmid was constructed that results in deletion of eleven COOH-terminal amino acids of human leukocyte A interferon and their replacement by nine unrelated amino acids encoded by the β-lactamase gene of pBR322. This altered human leukocyte interferon exhibits antiviral activity slightly lower than the natural molecule and appears to be more susceptible to proteolytic degradation as well. These results also confirm the conclusion that the eleven COOH-terminal amino acids are not essential for antiviral or antiproliferative activity.     

Mastoparan interacts with the carboxyl terminus of the alpha subunit of Gi

Mastoparan, a peptide toxin from wasp venom, stimulates guanine nucleotide binding and hydrolysis by G proteins. To elucidate the site of mastoparan-G protein interaction, we utilized a polyclonal antibody (R16,17) directed against the carboxyl terminus of the Gi alpha subunit to develop a competitive enzyme-linked immunosorbent assay. We investigated the ability of mastoparan to influence R16,17 antibody binding to G protein alpha subunits in a purified preparation of brain Gi and in neutrophil membrane extracts. Mastoparan antagonized the ability of R16,17 to detect G protein alpha subunits with an IC50 of 15 microM in the purified preparation and with an IC50 of 1 microM for the predominant G protein population in membrane extracts. This reduction was not seen when an unrelated peptide or a peptide of similar charge composition to mastoparan was used in place of mastoparan in the assay. Additionally, antibody R16,17 blocked up to 85% of mastoparan-stimulated GTPase activity. Taken together, these data indicate that the interaction of mastoparan with G protein depends in part on the carboxyl terminus of Gi alpha. Pertussis toxin-catalyzed ADP-ribosylation of Gi alpha markedly inhibited mastoparan-stimulated GTPase activity but only slightly attenuated the ability of mastoparan to recognize G protein. These data suggest that ribosylation inhibits mastoparan-induced G protein activation by a mechanism distinct from the ability of mastoparan to physically interact with G protein. Since mastoparan is thought to mimic hormone-liganded receptors, these findings may be applicable to the mechanism of receptor-Gi protein uncoupling that results from ADP-ribosylation of the G protein.     

A general method for preparation of peptides biotinylated at the carboxy terminus.

A method for the preparation of a biotinylated resin that can be elongated by standard methods of solid-phase peptide synthesis to give peptides biotinylated at the carboxy terminus is described. This methodology is particularly important for the preparation of biotinylated peptides in which a free amino terminus is required. Coupling of N epsilon-9-fluorenylmethoxycarbonyl-(Fmoc)-N alpha-tert-butyloxycarbonyl(Boc)-L- lysine to p-methylbenzhydrylamine resin, followed by removal of the Fmoc protecting group and reaction with (+)-biotin-4-nitrophenyl ester yielded N alpha-Boc-biocytin-p-methyl-benzhydrylamine resin. The utility of this resin was tested by the synthesis of a biotinylated peptide, Gly-Asn-Ala-Ala-Ala-Ala-Arg-Arg-biocytin-NH2, for use as an in vitro substrate for myristoyl-CoA:protein N-myristoyltransferase (NMT), the enzyme that catalyzes protein N-myristoylation. Analysis of the peptide derivative by HPLC and mass spectrometry revealed a single major product of the expected mass, indicating that the biotin group survived cleavage and deprotection with HF. The biotinylated peptide served as a substrate for NMT, and the resulting myristoylated peptide could be quantitatively recovered by adsorption to immobilized avidin.     

Antiserum specific for the carboxy terminus of the transforming protein of Rous sarcoma virus

An antiserum specific for the carboxy terminus of p60src, the transforming protein of Rous sarcoma virus, was produced by immunization of rabbits with a conjugate of bovine serum albumin and the synthetic peptide NH2-Tyr-Val-Leu-Glu-Val-Ala-Glu-COOH. The carboxy-terminal six amino acids of this peptide correspond in sequence to that deduced for the carboxy terminus of the p60src of the Schmidt-Ruppin strain of Rous sarcoma virus of subgroup A. The p60src proteins of the several strains of Rous sarcoma virus and the cellular homolog of the viral transforming protein, p60c-src, comprise a polymorphic family of polypeptides. The anticarboxy-terminal serum reacted readily with the p60src proteins of three different strains of Rous sarcoma virus. In contrast, no precipitation of cellular p60c-src could be detected with this serum. This suggests that the viral p60src proteins have identical carboxy termini and that the carboxy terminus of cellular p60c-src may be different from that of viral p60src. The anticarboxy-terminal serum reacted poorly with the subpopulation of viral p60src which is present in a complex with two cellular phosphoproteins. Apparently, the presence of the two cellular proteins interferes with the recognition of p60src by the anticarboxy-terminal serum. It seems likely, therefore, that these two cellular proteins bind to the carboxy-terminal domain of p60src.     

Effects of truncation of the carboxy terminus of pseudorabies virus glycoprotein B on infectivity

Glycoproteins homologous to the type I membrane glycoprotein B (gB) of herpes simplex virus 1 (HSV-1) are the most highly conserved glycoproteins within the family Herpesviridae and are present in members of each herpesvirus subfamily. In the alphaherpesvirus pseudorabies virus (PrV), gB is required for entry into target cells and for direct viral cell-to-cell spread. These processes, though related, appear to be distinct, and thus it was interesting to analyze whether they require different functions of gB. To this end, we established cell lines stably expressing different carboxy-terminally truncated versions of PrV gB by deleting either (i) one predicted intracytoplasmic alpha-helical domain encompassing putative YQRL and dileucine internalization signals, (ii) two predicted intracytoplasmic alpha-helical domains, (iii) the complete intracytoplasmic domain, or (iv) the intracytoplasmic domain and the transmembrane anchor region. Confocal laser scanning microscopy showed that gB derivatives lacking at least the last 29 amino acids (aa) localize close to the plasma membrane, while the full-length protein accumulates in intracellular aggregations. Trans-complementation studies with a gB-deleted PrV (PrV-gB(-)) demonstrated that the 29-aa truncated form lacking the putative internalization signals and the C-terminal alpha-helical domain (gB-008) was efficiently incorporated into PrV-gB(-) virions and efficiently complemented infectivity and cell-to-cell spread. Moreover, gB-008 exhibited an enhanced fusogenic activity. In contrast, gB proteins lacking both alpha-helical domains (gB-007), the complete intracytoplasmic domain, or the intracytoplasmic domain and transmembrane anchor were only inefficiently or not at all incorporated into PrV-gB(-) virions and did not complement infectivity. However, gB-007 was able to mediate cell-to-cell spread of PrV-gB(-). Similar phenotypes were observed when virus recombinants expressing gB-008 or gB-007, respectively, instead of wild-type gB were isolated and analyzed. Thus, our data show that internalization of gB is not required for gB incorporation into virions nor for its function in either entry or cell-to-cell spread. Moreover, they indicate different requirements for gB in these membrane fusion processes.     

Role of the carboxy terminus of Escherichia coli FtsA in self-interaction and cell division

The role of the carboxy terminus of the Escherichia coli cell division protein FtsA in bacterial division has been studied by making a series of short sequential deletions spanning from residue 394 to 420. Deletions as short as 5 residues destroy the biological function of the protein. Residue W415 is essential for the localization of the protein into septal rings. Overexpression of the ftsA alleles harboring these deletions caused a coiled cell phenotype previously described for another carboxy-terminal mutation (Gayda et al., J. Bacteriol. 174:5362-5370, 1992), suggesting that an interaction of FtsA with itself might play a role in its function. The existence of such an interaction was demonstrated using the yeast two-hybrid system and a protein overlay assay. Even these short deletions are sufficient for impairing the interaction of the truncated FtsA forms with the wild-type protein in the yeast two-hybrid system. The existence of additional interactions between FtsA molecules, involving other domains, can be postulated from the interaction properties shown by the FtsA deletion mutant forms, because although unable to interact with the wild-type and with FtsADelta1, they can interact with themselves and cross-interact with each other. The secondary structures of an extensive deletion, FtsADelta27, and the wild-type protein are indistinguishable when analyzed by Fourier transform infrared spectroscopy, and moreover, FtsADelta27 retains the ability to bind ATP. These results indicate that deletion of the carboxy-terminal 27 residues does not alter substantially the structure of the protein and suggest that the loss of biological function of the carboxy-terminal deletion mutants might be related to the modification of their interacting properties.     

Identification of peroxisomal targeting signals located at the carboxy terminus of four peroxisomal proteins

As part of an effort to understand how proteins are imported into the peroxisome, we have sought to identify the peroxisomal targeting signals in four unrelated peroxisomal proteins: human catalase, rat hydratase:dehydrogenase, pig D-amino acid oxidase, and rat acyl-CoA oxidase. Using gene fusion experiments, we have identified a region of each protein that can direct heterologous proteins to peroxisomes. In each case, the peroxisomal targeting signal is contained at or near the carboxy terminus of the protein. For catalase, the peroxisomal targeting signal is located within the COOH-terminal 27 amino acids of the protein. For hydratase:dehydrogenase, D-amino acid oxidase, and acyl-CoA oxidase, the targeting signals are located within the carboxy-terminal 15, 14, and 15 amino acids, respectively. A tripeptide of the sequence Ser-Lys/His-Leu is present in each of these targeting signals as well as in the peroxisomal targeting signal identified in firefly luciferase (Gould, S.J., G.-A. Keller, and S. Subramani. 1987. J. Cell Biol. 105:2923-2931). When the peroxisomal targeting signal of the hydratase:dehydrogenase is mutated so that the Ser-Lys-Leu tripeptide is converted to Ser-Asn-Leu, it can no longer direct proteins to peroxisomes. We suggest that this tripeptide is an essential element of at least one class of peroxisomal targeting signals.     

Cooperativity between the beta-tubulin carboxy tail and the body of the molecule is required for microtubule function

Using Drosophila spermatogenesis as a model, we show that function of the beta-tubulin C-terminal tail (CTT) is not independent of the body of the molecule. For optimal microtubule function, the beta-tubulin CTT and body must match. beta2 is the only beta-tubulin used in meiosis and spermatid differentiation. beta1-tubulin is used in basal bodies, but beta1 cannot replace beta2. However, when beta1 is co-expressed with beta2, both beta-tubulins are equally incorporated into all microtubules, and males exhibit near wild type fertility. In contrast, co-expression of beta2beta1C and beta1beta2C, two reciprocal chimeric molecules with bodies and tails swapped, results in defects in meiosis, cytoskeletal microtubules, and axonemes; males produce few functional sperm and few or no progeny. In these experiments, all the same beta-tubulin parts are present, but unlike the co-assembled native beta-tubulins, the "trans" configuration of the co-assembled chimeras is poorly functional. Our data thus reveal essential intra-molecular interactions between the CTT and other parts of the beta-tubulin molecule, even though the CTT is a flexible surface feature of tubulin heterodimers and microtubules. In addition, we show that Drosophila sperm tail length depends on the total tubulin pool available for axoneme assembly and spermatid elongation. D. melanogaster and other Drosophila species have extraordinarily long sperm tails, the length of which is remarkably constant in wild type flies. We show that in males of experimental genotypes that express wild type tubulins but have half the amount of the normal tubulin pool size, sperm tails are substantially shorter than wild type.     

End-to-end annealing of plant microtubules by the p86 subunit of eukaryotic initiation factor-(iso)4F.

The p86 subunit of eukaryotic initiation factor-(iso)4F from wheat germ exhibits saturable and substoichiometric binding to maize microtubules, induces microtubule bundling in vitro, and is colocalized or closely associated with cortical microtubule bundles in maize root cells, indicating its function as a microtubule-associated protein (MAP). The effects of p86 on the growth of short, taxol-stabilized maize microtubules were investigated. Pure microtubules underwent a gradual length redistribution, an increase in mean length, and a decrease in number concentration consistent with an end-to-end annealing mechanism of microtubule growth. Saturating p86 enhanced the microtubule length distribution and produced significantly longer and fewer microtubules than the control, indicating a facilitation of annealing by p86. Confirmation of endwise annealing rather than of dynamic instability as the mechanism for microtubule growth was made using mammalian MAP2, which also promoted the redistribution of length, increase in mean length, and decrease in number concentration of taxol-stabilized maize microtubules. Enhancement of microtubule growth occurred concomitant with bundling by p86, indicating that an alignment of microtubules in bundles facilitated endwise annealing kinetics. The results demonstrate that nonfacile plant microtubules can spontaneously elongate by endwise annealing and that MAPs enhance the rate of annealing. The p86 subunit of eukaryotic initiation factor-(iso)4F may be an important regulator of microtubule dynamics in plant cells.     

Monoclonal antibodies specific for the carboxy terminus of simian virus 40 large T antigen

Three mouse hybridomas secreting antibodies against the undecapeptide Lys-Pro-Pro-Thr-Pro-Pro-Pro-Glu-Pro-Glu-Thr, corresponding to the carboxy terminus of simian virus 40 large T antigen, were isolated and cloned. A sensitive enzyme-linked immunosorbent assay was used to characterize the properties of the monoclonal antibodies. All three hybridomas, designated KT1, KT3, and KT4, produced antibodies that immunoprecipitated large T. The antibodies differed in their affinities for the peptide and for the native protein. Antibodies from KT3 precipitated large T better than those from KT1 or KT4. KT3 antibodies also had the highest affinity for the free peptide (5.2 X 10(6) M-1) as determined by radioimmunoassay; KT1 and KT4 antibodies had ca. 5- and 1,000-fold lower affinities, respectively. Inhibition studies with shorter peptides, overlapping the undecapeptide, revealed the approximate regions recognized by the different monoclonal antibodies. KT3 antibodies bound to a region within the carboxy-terminal six amino acids of large T. Antibodies from KT1 and KT4 reacted with sequences located further towards the amino terminus of the undecapeptide. Surprising results were obtained with KT4 antibodies. Their binding to the undecapeptide was completely inhibited by the undecapeptide itself or the carboxy-terminal hexapeptide. The carboxy-terminal pentamer, on the other hand, slightly enhanced binding, and the carboxy-terminal tetramer, Glu-Pro-Glu-Thr, was strongly stimulatory. A model for this effect is proposed. Using the enzyme-linked immunosorbent assay, we confirmed previous studies (W. Deppert and G. Walter, Virology 122:56-70, 1982) which found that antiserum against sodium dodecyl sulfate-denatured large T reacts strongly with the carboxy terminus of large T. By inhibition studies, we identified the approximate region within the undecapeptide recognized by anti-sodium dodecyl sulfate-denatured large T and compared this region with the region identified by antipeptide serum.     

Coronavirus spike glycoprotein, extended at the carboxy terminus with green fluorescent protein, is assembly competent

Due to the limited ultrastructural information about the coronavirion, little is known about the interactions acting at the interface between nucleocapsid and viral envelope. Knowing that subtle mutations in the carboxy-terminal endodomain of the M protein are already lethal, we have now probed the equivalent domain of the spike (S) protein by extending it terminally with a foreign sequence of 27 kDa: the green fluorescent protein (GFP). When expressed individually in murine cells, the S-GFP chimeric protein induced the formation of fluorescent syncytia, indicating that it was synthesized and folded properly, trimerized, and transported to the plasma membrane, where it exhibited the two key S protein functions, i.e., interaction with virus receptor molecules and membrane fusion. Incorporation into virus-like particles demonstrated the assembly competence of the chimeric spike protein. The wild-type S gene of mouse hepatitis coronavirus (MHV) was subsequently replaced by the chimeric construct through targeted recombination. A viable MHV-SGFP was obtained, infection by which could be visualized by the fluorescence induced. The efficiency of incorporation of the chimeric protein into particles was, however, reduced relative to that in wild-type particles which may explain, at least in part, the reduced infectivity produced by MHV-SGFP infection. We conclude that the incorporation of spikes carrying the large GFP moiety is apparently impaired by geometrical constraints and selected against during the assembly of virions. Probably due to this disadvantage, deletion mutants, having lost the foreign sequences, rapidly evolved and outcompeted the chimeric viruses during virus propagation. The fluorescent MHV-SGFP will now be a convenient tool to study coronaviral cell entry.