Molecular characterization of hCTR1, the human copper uptake protein

We have expressed hCTR1, the human copper transporter, in Sf9 cells using a baculovirus-mediated expression system, and we observed greatly enhanced copper uptake. Western blots showed that the protein is delivered to the plasma membrane, where it mediates saturable copper uptake with a K(m) of approximately 3.5 microm. We also expressed functional transporters where the N-linked glycosylation sites were substituted, and we provided evidence for the extracellular location of the amino terminus. Accessibility of amino-terminal FLAG epitope to antibody prior to permeabilization and of carboxyl-terminal FLAG only after permeabilization confirmed the extracellular location of the amino terminus and established the intracellular location of the carboxyl terminus. Tryptic digestion of hCTR1 occurred within the cytoplasmic loop and generated a 10-Da carboxyl-terminal peptide; cleavage was prevented by the presence of copper. hCTR1 mutants where Cys-161 and Cys-189, the two native cysteines, were replaced with serines also mediated copper uptake, indicating that neither cysteine residue was essential for transport. However, the mutants provided evidence that these residues may stabilize hCTR1 oligomerization. Western blots of hCTR1 in Sf9 cells showed expression levels 100-fold higher than in mammalian (HepG2) cells. The high level of functional expression and the low level of endogenous copper uptake will enable future structure-function analysis of this important protein.     

Human Copper Transporter 1 Lacking O-Linked Glycosylation Is Proteolytically Cleaved in a Rab9-positive Endosomal Compartment

The human copper transporter hCTR1 is a homotrimer composed of a plasma membrane protein of 190 amino acids that contains three transmembrane segments. The extracellular 65-amino acid amino terminus of hCTR1 contains both N-linked (at Asn¹⁵) and O-linked (at Thr²⁷) sites of glycosylation. If O-glycosylation at Thr²⁷ is prevented, hCTR1 is efficiently cleaved, removing ∼30 amino acids from the amino terminus. We have now investigated (i) the site of this cleavage, determining which peptide bonds are cleaved, (ii) the mechanism by which glycosylation prevents cleavage, and (iii) where in the cell the proteolytic cleavage takes place. Cleavage occurs in the sequence Ala-Ser-His-Ser-His (residues 29–33), which does not contain previously recognized protease cleavage sites. Using a series of hCTR1 mutants, we show that cleavage occurs preferentially between residues Ala²⁹–Ser³⁰–His³¹. We also show that the O-linked polysaccharide at Thr²⁷ blocks proteolysis due to its proximity to the cleavage site. Moving the cleavage site away from the Thr²⁷ polysaccharide by insertion of as few as 5 amino acids allows cleavage to occur in the presence of glycosylation. Imaging studies using immunofluorescence in fixed cells and a functional green fluorescent protein-tagged hCTR1 transporter in live cells showed that the cleaved peptide accumulates in punctate structures in the cytoplasm. These puncta overlap compartments were stained by Rab9, indicating that hCTR1 cleavage occurs in a late endosomal compartment prior to delivery of the transporter to the plasma membrane.Copper is acquired by eukaryotic cells through transporters in the plasma membrane known as CTR proteins (1). Copper is an essential enzymatic cofactor in numerous proteins, many of which perform electron transfer reactions in which the metal cycles (2, 3) between the redox states (Cu⁺ and Cu²⁺) (4). This readily occurring redox reaction can make copper ions toxic to cells through the generation of reactive oxygen species. The free copper concentration in cells is extremely low (less than 1 fmol), and there is essentially no free copper in serum. Hence, copper transporters receive copper from copper-binding substrates in the serum, translocate it across the membrane, and transfer it to intracellular chaperones for delivery to target proteins (5).Human copper transporter 1 (hCTR1)² and orthologous proteins throughout eukaryotes have three transmembrane segments (6, 7) and form homotrimeric, membrane complexes (8, 9) that carry out the high affinity transport of monovalent copper (see Fig. 1, inset). The human hCTR1 gene was discovered by its ability to complement Saccharomyces cerevisiae yCtr mutants, demonstrating that high affinity copper transport is a conserved function among the CTR1 proteins (10). The CTR1 proteins range in size from 200 to 400 amino acids (1, 11), but share methionine- and histidine-rich motifs in the extracellular amino terminus, as well as conserved sequences in transmembrane segments (1, 12).Open in a separate windowFIGURE 1.Extracellular amino terminus of hCTR1. Location of N- and O-linked glycosylation at Asn¹⁵ and Thr²⁷, and the end points of 3 truncation mutants in gray: H22, A29, and G34. In the absence of O-glycosylation at Thr²⁷, hCTR1 is efficiently cleaved between A29 and G34 (black triangles). Location of the FLAG epitope tag is shown. Inset shows the complete 190-amino acid hCTR1 protein, with extracellular NH₂ terminus, three membrane spanning domains, intracellular loop, and COOH-terminal tail. The 5 amino acids in which cleavage occurs are shown in black. Three hCTR1 polypeptides form a symmetrical trimer in the copper transporter (8, 9).Little is known about the details of the copper transport mechanism in CTR1 proteins. Mutational studies of hCTR1 have identified a number of residues important for copper transport (12–14), such as methionine residues within the extracellular amino terminus, and two transmembrane segments that were important for ⁶⁴Cu uptake in cultured cells (12). A study of hCTR1 mutants expressed in insect cells identified residues in or near the transmembrane domains that affect K_m and or V_max of ⁶⁴Cu uptake (14). These results and recent structural studies suggest that copper transits a pore lined by transmembrane segments two and three in the homotrimeric complex (8, 9). Another mechanism based on endocytosis and degradation of hCTR1 has also been proposed (15).Vertebrate CTR1 proteins are widely expressed, and may play other roles in addition to copper transport. Mice homozygous for mCtr1 knock-out alleles die during midgestation, which was thought to reflect an early requirement for copper transport during development. However, a recent study showed that xCTR1 was part of a fibroblast growth factor signaling complex in Xenopus embryos active in Ras/extracellular signal-regulated kinase (ERK) signaling. The signaling role, which affects embryonic development in Xenopus and ES cell differentiation in mammalian cells, appears to be independent of the copper transport activity of CTR1 (16).In previous structure/function studies of hCTR1 we found that the extracellular amino terminus of ∼65 amino acids is modified by N- and O-linked glycosylation at Asn¹⁵ and Thr²⁷, respectively (6, 17) (see Fig. 1). N-Linked glycans at Asn¹⁵ increase the predicted mass of the hCTR1 polypeptide by about 9 kDa. Removing N-linked polysaccharides by a N15Q mutation does not significantly affect the expression or function of the transporter (6, 17). O-Linked polysaccharides at Thr²⁷ that terminate in sialic acid residues increase the mass of the polypeptide by 1–2 kDa, (17). In the absence of O-linked glycosylation, the polypeptide undergoes very efficient cleavage near Thr²⁷, leaving a 17-kDa hCTR1 protein lacking about 30 amino acids from the extracellular amino terminus (Fig. 1). The truncated (17 kDa) hCTR1 protein was efficiently delivered to the plasma membrane, but exhibited only 50–60% of the copper transport activity of wild-type hCTR1 (17).In recent years, an impressive variety of proteases have been characterized in the secretory pathway and plasma membrane (18–22). Many of these proteases perform some kind of regulatory cleavage, from maturation of pre-proteins, (including proteases), to membrane proteases involved in shedding of ectodomains. Presumably, cleavage of hCTR1 lacking O-linked glycosylation must occur after the addition of the O-linked sugars would have occurred in the golgi (23). Cleavage of the unglycosylated hCTR1 protein could thus occur while the transporter is en route to the plasma membrane (23, 24), after delivery to the surface, or, as in the case of some receptors, during recycling between the plasma membrane and interior compartments (25–28).In this report, we show that inhibition of cleavage by O-linked glycosylation at Thr²⁷ requires close proximity of the polysaccharide to the site of cleavage. Moving the cleavage site away from Thr²⁷ polysaccharides allowed cleavage. In mutants lacking O-glycosylation, hCTR1 is cleaved within amino acids 29–33 (ASHSH), preferentially between Ala²⁹–Ser³⁰–His³¹. Live cell imaging of GFP-tagged mutant hCTR1 and staining of fixed cells overexpressing FLAG-tagged hCTR1 shows that the cleaved amino-terminal peptides accumulate in punctate structures that partially overlap Rab9, a late endosome marker, suggesting that cleavage occurs after transit through the golgi, but prior to delivery to the plasma membrane.     

Basic amino acids as modulators of an O-linked glycosylation signal of the herpes simplex virus type 1 glycoprotein gC: functional roles in viral infectivity

The herpes simplex virus type 1 (HSV-1) glycoprotein gC-1 is engaged both in viral attachment and viral immune evasion mechanisms in the infected host. Besides several N-linked glycans, gC-1 contains numerous O-linked glycans, mainly localized in two pronase-resistant clusters in the N-terminal domain of gC-1. In the present study we construct and characterize one gC-1 mutant virus, in which two basic amino acids (114K and 117R) in a putative O-glycosylation sequon were changed to alanine. We found that this modification did not modify the N-linked glycosylation but increased the content of O-linked glycans considerably. Analysis of the O-glycosylation capacity of wild-type and mutant gC-1 was performed by in vitro glycosylation assays with synthetic peptides derived from the mutant region predicted to present new O-glycosylation sites. Thus the mutant peptide region served as a better substrate for polypeptide GalNAc-transferase 2 than the wild-type peptide, resulting in increased rate and number of O-glycan attachment sites. The predicted increase in O-linked glycosylation resulted in two modifications of the biological properties of mutant virus-that is, an impaired binding to cells expressing chondroitin sulfate but not heparan sulfate on the cell surface and a significantly reduced plaque size in cultured cells. The results suggested that basic amino acids present within O-glycosylation signals may down-regulate the amount of O-linked glycans attached to a protein and that substitution of such amino acid residues may have functional consequences for a viral glycoprotein involving virus attachment to permissive cells as well as viral cell-to-cell spread.     

COOH-terminal processing of polypeptide D1 of the photosystem II reaction center of Scenedesmus obliquus is necessary for the assembly of the oxygen-evolving complex

Mutant LF-1 of the green alga Scenedesmus obliquus has been described by Metz and co-workers (Metz, J. G., Pakrasi, H., Seibert, M., and Arntzen, C. J. (1986) FEBS Lett. 205, 269-274) to be inactive for light-driven oxygen evolution, despite a functional Photo-system II reaction center. A polypeptide, D1, implicated in the ligation of the primary photoreactants of photosystem II, was shown to migrate with an apparent higher molecular mass on LDS-PAGE in the mutant than in the wild-type (WT) strain. We show here that polypeptide D1 is synthesized in a precursor form in Scenedesmus WT. Following synthesis and insertion into the thylakoid membrane, a 1.5-2-kDa oligopeptide is clipped off with a half-time of 1-2 min, yielding the mature 34-kDa form of the polypeptide. No processing of polypeptide D1 from mutant LF-1 was observed to take place. We show here that polypeptide D1 of LF-1 displays an identical proteolytic fingerprint pattern to the precursor D1 polypeptide of the wild-type strain. These both have molecular masses about 1.5-2 kDa higher than that of the mature WT polypeptide. A polyclonal antibody elicited by a synthetic oligopeptide (14-mer), predicted from the psbA gene nucleotide sequence to be homologous to the COOH terminus of the precursor D1 of spinach, cross-reacts only with D1 of mutant LF-1 and not with mature D1 of spinach, Chlamydomonas, or of Scenedesmus WT. This observation demonstrates that the greater molecular mass of polypeptide D1 from mutant LF-1 and of Scenedesmus WT precursor D1 is derived from a COOH-terminal extension. We conclude that the LF-1 mutant lacks the appropriate nuclear-encoded protease which processes polypeptide D1 at its COOH terminus from the precursor to the mature form. Such processing would appear to be a necessary step toward the stable incorporation of manganese into the oxygen-evolving site.     

Structure and cell surface maturation of the attachment glycoprotein of human respiratory syncytial virus in a cell line deficient in O glycosylation.

The synthesis of the extensively O-glycosylated attachment protein, G, of human respiratory syncytial virus and its expression on the cell surface were examined in a mutant Chinese hamster ovary (CHO) cell line, ldlD, which has a defect in protein O glycosylation. These cells, used in conjunction with an inhibitor of N-linked oligosaccharide synthesis, can be used to establish conditions in which no carbohydrate addition occurs or in which either N-linked or O-linked carbohydrate addition occurs exclusively. A recombinant vaccinia virus expression vector for the G protein was constructed which, as well as containing the human respiratory syncytial virus G gene, contained a portion of the cowpox virus genome that circumvents the normal host range restriction of vaccinia virus in CHO cells. The recombinant vector expressed high levels of G protein in both mutant ldlD and wild-type CHO cells. Several immature forms of the G protein were identified that contained exclusively N-linked or O-linked oligosaccharide side chains. Metabolic pulse-chase studies indicated that the pathway of maturation for the G protein proceeds from synthesis of the 32-kilodalton (kDa) polypeptide accompanied by cotranslational attachment of high-mannose N-linked sugars to form an intermediate with an apparent mass of 45 kDa. This step is followed by the Golgi-associated conversion of the N-linked sugars to the complex type and the completion of the O-linked oligosaccharides to achieve the mature 90-kDa form of G. Maturation from the 45-kDa N-linked form to the mature 90-kDa form occurred only in the presence of O-linked sugar addition, confirming that O-linked oligosaccharides constitute a significant proportion of the mass of the mature G protein. In the absence of O glycosylation, forms of G bearing galactose-deficient truncated N-linked and fully mature N-linked oligosaccharides were observed. The effects of N- and O-linked sugar addition on the transport of G to the cell surface were measured. Indirect immunofluorescence and flow cytometry showed that G protein could be expressed on the cell surface in the absence of either O glycosylation or N glycosylation. However, cell surface expression of G lacking both N- and O-linked oligosaccharides was severely depressed.     

Proteolytic degradation of ferredoxin-NADP reductase during purification from spinach

Ferredoxin-NADP reductase (FNR) was rapidly isolated from spinach leaves with special care to suppress proteolytic degradation. The molecular mass of this FNR preparation was estimated to be 35 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Limited proteolysis of 35-kDa FNR to 33-kDa FNR was effectively suppressed by high pH (at pH 9.3), concentrated salts, and low temperature. On the basis of these observations, a new isolation procedure was designed to obtain 35-kDa FNR in a preparative scale. The resulting final preparation still contained two FNR components. One appeared to correspond to the longest polypeptide so far reported for spinach FNR (Karplus et al., 1984, Biochemistry 23, 6576-6583) while the other lacked a gamma-pyroglutamyl residue from its amino terminus. Conventional preparation procedure without suppression of proteolytic action yielded an FNR preparation with a molecular mass of 33 kDa. This FNR preparation consisted of three components. They lacked 11 to 17 amino-terminal residues, while their carboxyl-terminal structure was retained intact. These results showed that proteolytic degradation of the spinach FNR molecule during purification took place exclusively at its amino-terminal moiety and further suggested that 35-kDa FNR with Karplus' structure should be the mature FNR molecule functional in the chloroplast thylakoids.     

The sodium ion translocating oxalacetate decarboxylase of Klebsiella pneumoniae. Sequence of the biotin-containing alpha-subunit and relationship to other biotin-containing enzymes

The gene encoding the alpha-subunit of the Na+ pump oxalacetate decarboxylase of Klebsiella pneumoniae was cloned and sequenced. The deduced primary structure of the protein was confirmed by protein sequencing of about 30% of the polypeptide chain. The gene has a GC content of 67% and codes for 596 amino acids. The N-terminal methionine is removed in the mature protein which has a calculated molecular mass of 63,600 daltons. The protein consists of two different domains that are connected by a stretch of amino acid residues susceptible to proteolytic cleavage. Limited proteolysis of the native enzyme with trypsin produced fragments of about 51 kDa and 10.2 kDa, the latter of which started with valine 491 and contained the biotin prosthetic group. Peptide sequencing indicated binding of the biotin prosthetic group to lysine 561, 35 residues from the C terminus. The decarboxylase contains an extended alanine- and proline-rich region (positions 502-532) on the N-terminal side of the 10.2-kDa biotin domain. This sequence includes a total of 16 alanine and 9 proline residues.     

Structural characterization of human interferon gamma. Heterogeneity of the carboxyl terminus

Natural human interferon gamma(IFN-gamma) was purified from the conditioned medium of peripheral blood leukocytes using selective silica gel adsorption and antibody-affinity chromatography. SDS-PAGE and Western blot analysis demonstrated three major species with molecular masses of 25 kDa, 20 kDa and 17 kDa. Structural analysis of this natural IFN-gamma preparation demonstrated a pyroglutamate residue at the amino terminus and a heterogeneous carboxyl terminus. The longest and most predominant polypeptide was 138 amino acids in length, which is five residues shorter than the sequence predicted from the cDNA. The presence of multiple-carboxyl-terminal forms indicated possible proteolytic processing during induction or protein purification. Limited proteolytic digestion of full-length recombinant IFN-gamma with endoproteinase Lys-C and trypsin revealed that the carboxyl-terminal 15 residues could be released under conditions in which the core portion of the polypeptide chain remained intact. Thus, the heterogeneity of natural IFN-gamma may be explained by partial proteolytic degradation of the molecule and differences in the degree of glycosylation as previously reported [Rinderknecht, E., O'Conner, B. H. & Rodriguez, H. (1984) J. Biol. Chem. 259, 6790-6797].     

Structural properties of a soluble bioactive precursor for transforming growth factor-alpha

The precursor for transforming growth factor-alpha, proTGF-alpha, is synthesized as an integral membrane glycoprotein with the mature TGF-alpha sequence located in the extracellular domain. Retrovirally transformed rat embryo fibroblasts (FeSV-Fre cells) expressing the endogenous proTGF-alpha gene release and accumulate in the medium mature TGF-alpha as well as a heterogeneous (17-19 kDa) group of soluble, bioactive TGF-alpha precursor forms. These precursors correspond to the heterogeneously glycosylated extracellular domain of proTGF-alpha which is released from the membrane by proteolytic cleavage. They are designated mesoTGF-alpha to denote their intermediate position in the proTGF-alpha processing pathway. The nature of the carbohydrate linked to mesoTGF-alpha has been examined by treatment with glycosidases and the use of metabolic inhibitors of glycosylation. The results indicate that the TGF-alpha precursors from FeSV-Fre cells contain O-linked carbohydrate as well as sialylated N-linked carbohydrate. Heterogeneous N-linked glycosylation of an 11-kDa core polypeptide accounts for the heterogeneous nature of mesoTGF-alpha. MesoTGF-alpha released by cells treated with inhibitors of N-linked carbohydrate processing appears as a 17-kDa species. Treatment with these inhibitors does not alter significantly the production of mesoTGF-alpha or mature TGF-alpha by the cells. However, treatment of cells with an inhibitor of co-translational N-linked glycosylation, tunicamycin, reduces the accumulation of mesoTGF-alpha in the medium and blocks the production of mature TGF-alpha under conditions in which overall protein synthesis is only minimally affected. These findings suggest that the proTGF-alpha processing activity is limiting in FeSV-Fre cells and other transformed cells that accumulate mesoTGF-alpha in the medium and that proTGF-alpha processing depends on a component whose function may require N-linked glycosylation.     

O-Glycosylation of the V2 vasopressin receptor.

The human V2 vasopressin receptor contains one consensus site for N-linked glycosylation at asparagine 22 in the predicted extracellular amino terminal segment of the protein. This segment also contains clusters of serines and threonines that are potential sites for O-glycosylation. Mutagenesis of asparagine 22 to glutamine abolished N-linked glycosylation of the V2 receptor (N22Q-V2R), without altering its function or level of expression. The N22Q-V2R expressed in transfected cells migrated in denaturing acrylamide gels as two protein bands with a difference of 7000 Da. Protein labeling experiments demonstrated that the faster band could be chase to the slower one suggesting the presence of O-linked sugars. Sialidase treatment of membranes from cells expressing the N22Q-V2R or of immunoprecipitated metabolically labeled V2R accelerated the migration of the protein in acrylamide gels demonstrating the existence of O-glycosylation, the first time this type of glycosylation has been found in a G protein coupled receptor. Synthesis of metabolically labeled receptor in the presence of 1 mM phenyl-N-acetyl-alpha-D-galactosaminide, a competitive inhibitor of N-acetyl-alpha-D-galactose and N-acetylneuraminic acid transferases, also produced a receptor that migrated faster in denaturing gels. Serines and threonines present in the amino terminus were analyzed by alanine scanning mutagenesis to identify the acceptor sites. O-glycosylation was found at most serines and threonines present in the amino terminus. Because the disappearance of a site opened the availability of others to the transferases, the exact identification of the acceptor sites was not feasible. The wild type V2R expressed in HEK 293, COS, or MDCK cells underwent N- and O-linked glycosylation. The mutant V2R bearing all serine/threonine substitutions by alanine at the amino terminus yielded a receptor functionally indistinguishable from the wild type protein, whose mobility in polyacrylamide gels was no longer affected by sialidase treatment.     

Competence for natural transformation in Neisseria gonorrhoeae: components of DNA binding and uptake linked to type IV pilus expression

Catalytically active, recombinant fusion proteins of bacteriophage E endosialidase were expressed and purified from Escherichia coli. Constructs with different fusion partners added to the amino terminus of the endosialidase were enzymatically active. A post-translational proteolytic cleavage was shown to occur between serine 706 and aspartate 707 to generate the 76 kDa mature enzyme from the 90 kDa translation product. Endosialidase truncated at the C-terminus from aspartate 707 was observed to have the same 76 kDa molecular weight as wild-type enzyme using denaturing SDS-PAGE but, under native PAGE conditions, was not observed to form the approximately 250 kDa trimeric wild-type enzyme, implying that the C-terminus of the enzyme may be required for correct assembly of active trimer, rather than as part of the active site as has been previously suggested. Mutagenesis of aspartate 138 to alanine greatly reduced enzyme activity whereas conversion of other selected aspartate residues to alanine had less effect, consistent with similarities between the structure and cata-lytic mechanism of bacteriophage E endosialidase and those of exosialidases.     

IgE receptor on human lymphocytes. IV. Further analysis of its structure and of the role of N-linked carbohydrates

Previous studies have shown that the low affinity receptor for IgE (Fc epsilon R II) on human B lymphocytes was comprised of three components with apparent Mr of 45, 65 to 95, and 37 kDa. The present results indicate that the 37-kDa component is a soluble degradation fragment of the 45-kDa component and they also suggest that the 65- to 95-kDa component is probably made of aggregates of the above components that are formed after solubilization of the cells. The 45-kDa component appears to be a glycoprotein containing several sialic acid residues, O-linked carbohydrates and one N-linked carbohydrate chain that is of the complex type. Partial digestion of the purified 65- to 95-, 45-, and 37-kDa molecules with alpha-chymotrypsin or protease V8 generates several fragments with identical mobility on SDS-PAGE. The 37 kDa is not N-glycosylated but like the IgE-binding factors present in the culture supernatant of Fc epsilon R-bearing cells, it contains sialic acid and O-linked carbohydrates. On incubation with protease inhibitors, the Mr of IgE-binding factors (BF) is shifted from 25-27 to 37 kDa, indicating that IgE-BF are derived from the proteolytic cleavage of the 37-kDa molecule, previously identified as a membrane component of Fc epsilon R. On incubation with N-glycosylation inhibitors, the production of IgE-BF is significantly increased indicating that N-glycosylation inhibits the degradation of Fc epsilon R into IgE-BF. Inasmuch as the effect of glycosylation inhibitors is not prevented by monensin, it is concluded that all the IgE-BF are derived from surface Fc epsilon R and not from their intracellular precursors.     

Localization of the sites of ADP-ribosylation and GTP binding in the eukaryotic elongation factor EF-2

Tryptic cleavage of EF-2, molecular mass 93 kDa, produced an 82-kDa polypeptide and a 10-kDa fragment, which was further degraded. By a slower reaction the 82-kDa polypeptide was gradually split into a 48-kDa and a 34-kDa fragment. Similarly, treatment with chymotrypsin resulted in the formation of an 82-kDa polypeptide and a small fragment. In contrast to the tryptic 82-kDa polypeptide the corresponding chymotryptic cleavage product was relatively resistant to further attack. The degradation of the 82-kDa polypeptide with either trypsin or chymotrypsin was facilitated by the presence of guanosine nucleotides, indicating a conformational shift in native EF-2 upon nucleotide binding. No effect was observed in the presence of ATP, indicating that the effect was specific for guanosine nucleotides. After affinity labelling of native EF-2 with oxidized [3H]GTP and subsequent trypsin treatment the radioactivity was recovered in the 48-kDa polypeptide showing that the GTP-binding site was located within this part of the factor. Correspondingly, tryptic degradation of EF-2 labelled with [14C]NAD+ in the presence of diphtheria toxin showed that the site of ADP-ribosylation was within the 34-kDa polypeptide. By cleavage with the tryptophan-specific reagent N-chlorosuccinimide the site of ADP-ribosylation could be located at a distance of 40-60 kDa from the GTP-binding site and about 4-11 kDa from the nearest terminus.     

Proteolytic processing yields two secreted forms of sonic hedgehog.

Sonic hedgehog (Shh) is expressed in tissues with known signalling capacities, such as the notochord, the floor plate of the central nervous system, and the zone of polarizing activity in the limb. Several lines of evidence indicate that Shh is involved in floor plate induction, somite patterning, and regulation of anterior-posterior polarity in the vertebrate limb. In this report, we investigate the biochemical behavior of Shh in a variety of expression systems and embryonic tissues. Expression of mouse Shh in Xenopus oocytes, COS cells, and baculovirus-infected insect cells demonstrates that in addition to signal peptide cleavage and N-linked glycosylation, chicken and mouse Shh proteins undergo additional proteolytic processing to yield two peptides with molecular masses of approximately 19 kDa (amino terminus) and 27 kDa (carboxy terminus), both of which are secreted. In transfected COS cells, we show that the 19-kDa peptide does not accumulate significantly in the medium unless heparin or suramin is added, suggesting that this peptide associates with the cell surface or extracellular matrix. This retention appears to depend on sequences in the carboxy-terminal part of the peptide. Finally, detection of the 19-kDa product in a variety of mouse and chicken embryonic tissues demonstrates that the proteolytic processing observed in cell culture is a normal aspect of Shh processing in embryonic development. These results raise the possibility that amino- and carboxyl-terminal regions of Shh may have distinct functions in regulating cell-cell interactions in the vertebrate embryo.     

A new structural model for the thyrotropin (TSH) receptor, as determined by covalent cross-linking of TSH to the recombinant receptor in intact cells: evidence for a single polypeptide chain.

The most widely held model for the human TSH receptor is of holoreceptor of 80 kDa with two subunits of approximately 50 and 30 kDa linked by disulfide bridges, with the former subunit containing the major hormone-binding site. We reexamined this model by covalently cross-linking radiolabeled TSH to the recombinant human TSH receptor stably expressed in Chinese hamster ovary (CHO) cells. When cross-linking was performed after the preparation of CHO membranes, analysis of hormone-receptor complexes under reducing and nonreducing conditions provided results supporting the two-subunit TSH receptor model. In contrast, however, cross-linking of TSH to the TSH receptor in intact CHO cells before membrane preparation revealed, even under reducing conditions, an approximately 100-kDa receptor as well as an approximately 54-kDa hormone-binding subunit. The approximately 100-kDa holoreceptor size is consistent with the size of the TSH receptor, as predicted from its derived amino acid sequence. The proportions of the approximately 100-kDa TSH receptor and the 54-kDa fragment varied in different experiments, suggesting the occurrence of proteolytic cleavage. Cross-linking of radiolabeled TSH to intact cells expressing a mutant TSH receptor (TSHR-D1) lacking amino acids 317-366 localized the proteolytic cleavage site to just up-stream of amino acid residue 317. In summary, the present data obtained by cross-linking TSH to recombinant human TSH receptors in intact cells provides evidence that the receptor exists in vivo as an approximately 100-kDa glycoprotein with a single polypeptide chain with intramolecular disulfide bridges.(ABSTRACT TRUNCATED AT 250 WORDS)     

Secretion of N-glycosylated interleukin-1 beta in Saccharomyces cerevisiae using a leader peptide from Candida albicans. Effect of N-linked glycosylation on biological activity

Human interleukin-1 beta (IL-1 beta) is expressed in activated monocytes as a 31-kDa precursor protein which is processed and secreted as a mature, unglycosylated 17-kDa carboxyl-terminal fragment, despite the fact that it contains a potential N-linked glycosylation site near the NH2 terminus (-Asn7-Cys8-Thr9-). cDNA coding for authentic mature IL-1 beta was fused to the signal sequence from the Candida albicans glucoamylase gene, two amino acids downstream from the signal processing site. Upon expression in Saccharomyces cerevisiae, approximately equimolar amounts of N-glycosylated (22 kDa) and unglycosylated (17 kDa) IL-1 beta protein were secreted. The N-glycosylated yeast recombinant IL-1 beta exhibited a 5-7-fold lower specific activity compared to the unglycosylated species. The mechanism responsible for inefficient glycosylation was also studied. We found no differences in secretion kinetics or processing between the two extracellular forms of IL-1 beta. The 17-kDa protein, which was found to lack core sugars, does not result from deglycosylation of the 22-kDa protein in vivo and does not result from saturation of the glycosylation enzymatic machinery through overexpression. Alteration of the uncommon Cys8 residue in the -Asn-X-Ser/Thr-glycosylation site to Ser also had no effect. However, increasing the distance between Asn7 and the signal processing site increased the extent of core N-linked glycosylation, suggesting a reduction in glycosylation efficiency near the NH2 terminus.     

A critical proteolytic cleavage site near the C terminus of the yeast retrotransposon Ty1 Gag protein.

Cleavage of the Gag and Gag-Pol polyprotein precursors is a critical step in proliferation of retroviruses and retroelements. The Ty1 retroelement of Saccharomyces cerevisiae forms virus-like particles (VLPs) made of the Gag protein. Ty1 Gag is not obviously homologous to the Gag proteins of retroviruses. The apparent molecular mass of Gag is reduced from 58 to 54 kDa during particle maturation. Antibodies raised against the C-terminal peptide of Gag react with the 58-kDa polypeptide but not with the 54-kDa one, indicating that Gag is proteolytically processed at the C terminus. A protease cleavage site between positions 401 and 402 of the Gag precursor was defined by carboxy-terminal sequencing of the processed form of Gag. Certain deletion and substitution mutations in the C terminus of the Gag precursor result in particles that are two-thirds the diameter of the wild-type VLPs. While the Ty1 protease is active in these mutants, their transposition rates are decreased 20-fold compared with that of wild-type Ty1. Thus, the Gag C-terminal portion, released in the course of particle maturation, probably plays a significant role in VLP morphogenesis and Ty1 transposition.     

Construction by site-directed mutagenesis of a 39-kilodalton mosquitocidal protein similar to the larva-processed toxin of Bacillus sphaericus 2362.

After ingestion of the parasporal crystals of Bacillus sphaericus, mosquito larvae process the 42-kilodalton (kDa) toxin to a protein of 39 kDa, which has an increased toxicity (A. H. Broadwell and P. Baumann, Appl. Environ. Microbiol. 53:1333-1337, 1987). A similar activation is performed by trypsin and chymotrypsin. Using site-directed mutagenesis, we have constructed derivatives of the 42-kDa toxin with a deletion of 10 amino acids at the N terminus and deletions of 7, 17, or 20 amino acids at the C terminus. Toxicity for mosquito larvae was retained upon deletion of 7 or 17 amino acids but was lost upon deletion of 20 amino acids. Evidence is presented indicating that the protein containing deletions of 10 amino acids at the N terminus and 17 amino acids at the C terminus (corresponding to potential chymotrypsin cleavage sites) is similar to the 39-kDa protein produced in mosquito larvae or by digestion with chymotrypsin. Digestion with trypsin appears to generate a protein lacking 16 or 19 amino acids from the N terminus and 7 amino acids from the C terminus. As is the case with the recombinant-made 42-kDa protein, toxicity of its derivatives is dependent on the presence of a 51-kDa protein which is a component of the parasporal crystal of B. sphaericus 2362.