Differential degradation of the three fibrinogen chains by proteasomes: involvement of Sec61p and cytosolic Hsp70.

HepG2 cells, which synthesize and secrete fibrinogen, accumulate surplus Aalpha and gamma chains. The nonsecreted fibrinogen chains are degraded both by proteasomes and lysosomes, with unassembled chains primarily degraded by proteasomes and an Aalpha-gamma complex by lysosomes. To further determine the mechanisms by which unassembled fibrinogen chains are degraded, and to explain the pools of Aalpha and gamma chains that occur in HepG2 cells, the association of fibrinogen chains with Sec61beta, a component of the translocon, and with a cytosol chaperone, Hsp70, was studied in both HepG2 cells and COS cells expressing single fibrinogen chains. Retrotranslocation from the lumen of the endoplasmic reticulum was shown by treatment with MG132, a proteasome inhibitor. MG132 caused glycosylated Bbeta to accumulate on Sec61beta in COS cells expressing Bbeta and acted similarly with all three fibrinogen chains in HepG2 cells. In HepG2 cells, Bbeta was associated with Sec61beta ahead of Aalpha and gamma chains, suggesting that pools of Aalpha and gamma chains may be caused by unequal rates of retrotranslocation. In COS cells, retrotranslocation into the cytoplasm was demonstrated by the ATP-sensitive association of ubiquitinylated Aalpha, Bbeta, and gamma chains bound to Hsp70. More Aalpha and gamma than Bbeta accumulated on Hsp70 of HepG2 cells, consistent with more rapid degradation of Bbeta. Overexpression of Hsp70 in HepG2 cells resulted in decreased secretion, but not synthesis, of fibrinogen. Decreased secretion may be due to enhanced degradation of unassembled fibrinogen chains, indicating that proteolysis by proteasomes might regulate both the intracellular pools of fibrinogen chains and fibrinogen secretion.

First example of anti-Kx in a person with the McLeod phenotype and without chronic granulomatous disease.

BACKGROUND: Kx is lacking in the RBCs of patients with the McLeod syndrome. This condition is sometimes associated with chronic granulomatous disease (CGD). If given allogeneic RBCs, CGD patients with the McLeod phenotype may produce anti-Kx and anti-Km, and only phenotypically matched McLeod blood would be compatible. McLeod phenotype persons without CGD have made anti-Km but not anti-Kx (2 examples), and thus both McLeod and K(O) blood would be compatible. CASE REPORT: RBCs from a transfused patient with the McLeod phenotype but not with CGD (non-CGD McLeod) were typed for the Kell blood group antigens, and the plasma was analyzed for the presence of antibody by agglutination. The molecular basis was determined by analyzing for XK protein on RBC membranes by Western immunoblotting, by sequencing the XK gene, and by RFLP. RESULTS: The RBCs did not react with anti-Kx + anti-Km and showed weakening of Kell system antigens. The patient's plasma reacted moderately (2+) with RBCs of common Kell type and strongly (4+) with K(O) RBCs and RBCs of common Kell type treated with dithiothreitol, and did not react with McLeod RBCs. XK protein was absent from the RBC membranes. The XK gene had a point mutation in the donor splice site of intron 1 (G>C). CONCLUSION: This is the first report describing the molecular alteration in a non-CGD McLeod patient who has made anti-Kx. The immune response of people with the McLeod phenotype can vary, and K(O) blood may not always be compatible.

The mouse Kell blood group gene (Kel): cDNA sequence, genomic organization, expression, and enzymatic function.

The human Kell blood group system is important in transfusion medicine, since Kell is a polymorphic protein and some of its antigens can cause severe reactions if mismatched blood is transfused, while maternal alloimmunization may lead to fetal and neonatal anemia. In humans, Kell is an Mr 93,000 type II membrane glycoprotein with endothelin-3-converting enzyme activity that is linked by a single disulfide bond to another protein, XK, that spans the membrane ten times. An absence of XK leads to clinical symptoms termed the McLeod syndrome. We determined the cDNA sequence of the mouse Kell homologue, the organization of the gene, expression of the protein and its enzymatic function on red cells. Comparison of human and mouse Kell cDNA showed 80% nucleotide and 74% amino acid sequence identity. Notable differences are that the mouse Kell protein has eight probable N-linked carbohydrate side chains, compared to five for human Kell, and that the mouse homologue has one more extracellular cysteine than human Kell protein. The mouse Kell gene (Kel), like its human counterpart, is similarly organized into 19 exons. Kel was located to proximal Chromosome 6. Northern blot analysis showed high expression in spleen and weaker levels in testis and heart. Western blot analysis of red cell membrane proteins demonstrated that mouse Kell glycoprotein has an apparent Mr of 110,000 and, on removal of N-linked sugars, 80,000. As in human red cells, Kell is disulfide-linked to XK and mouse red cells have endothelin-3-converting enzyme activity.

The Kell blood group system: Kell and XK membrane proteins.

Two membrane proteins express the antigens that comprise the Kell blood group system. A single antigen, Kx, is carried on XK, a 440-amino acid protein that spans the membrane 10 times, and more than 20 antigens reside on Kell, a 93-kd, type II glycoprotein. XK and Kell are linked, close to the membrane surface, by a single disulfide bond between Kell cysteine 72 and XK cysteine 347. Although primarily expressed in erythroid tissues, Kell and XK are also present in many other tissues. The polymorphic forms of Kell are due to single base mutations that encode different amino acids. Some Kell antigens are highly immunogenic and may cause strong reactions if mismatched blood is transfused and severe fetal anemia in sensitized mothers. Antibodies to KEL1 may suppress erythropoiesis at the progenitor level, leading to fetal anemia. The cellular functions of Kell/XK are complex. Absence of XK, the McLeod phenotype, is associated with acanthocytic red blood cells (RBCs), and with late-onset forms of muscular dystrophy and nerve abnormalities. Kell, by homology, is a member of the neprilysin (M13) family of membrane zinc endopeptidases and it preferentially activates endothelin-3 by specific cleavage of the Trp21-Ile22 bond of big endothelin-3.

Kell, Kx and the McLeod syndrome.

The antigens of the Kell blood group system are carried on a 93 kDa type II glycoprotein encoded by a single gene on chromosome 7 at 7q33. XK is a 50.9 kDa protein that traverses the membrane ten times and derives from a single gene on the X chromosome at Xp21. A single disulphide bond, Kell Cys 72-XK Cys 347, links Kell to XK. The Kell component of the Kell/XK complex is important in transfusion medicine since it is a highly polymorphic protein, carrying over 23 different antigens, that can cause severe reactions if mismatched blood is transfused and in pregnant mothers antibodies to Kell may elicit serious fetal and neonatal anaemia. The different Kell phenotypes are all caused by base mutations leading to single amino acid substitutions. By contrast the XK component carries a single blood group antigen, termed Kx. The physiological functions of Kell and XK have not been fully elucidated but Kell is a zinc endopeptidase with endothelin-3-converting enzyme activity and XK has the structural characteristics of a membrane transporter. Lack of Kx, the McLeod phenotype, is associated with red cell acanthocytosis, elevated levels of serum creatine phosphokinase and late onset forms of muscular and neurological defects.

The EC domains of human fibrinogen420 contain calcium binding sites but lack polymerization pockets.

The extended (E) isoform unique to Fibrinogen420 (Fib420) is distinguished from the conventional chain of Fibrinogen340 by the presence of an additional 236-residue carboxyl terminus globular domain (EC). A recombinant form of EC (rEC), having a predicted mass of 27,653 Daltons, was expressed in yeast (Pichia pastoris) and purified by anion exchange column chromatography. Purified rEC appears to be predominantly intact, as judged by N-terminal sequence analysis, mass spectral analysis of the C-terminal cyanogen bromide (CNBr) fragment, and comparison of recognition by epitope-specific monoclonal antibodies. Carbohydrate determination, coupled with analysis of CNBr digestion fragments, confirms N-linked glycosylation at Asn667, the site at which sugar is attached in E. Analysis of CNBr digestion fragments confirms that two disulfide bridges exist at cysteine pairs E613/644 and E780/793. In the presence of 5 mmol/L EDTA, rEC is highly susceptible to plasmic degradation, but Ca2+ (5 mmol/L) renders rEC resistant. No protective effect from plasmic degradation was conferred to rEC by the peptides GPRPamide or GHRP, nor did rEC bind to a GPR peptide column. These results suggest that the EC domain contains a calcium-binding site, but lacks a polymerization pocket. By analogy with the site elucidated in the gammaC domain, we predict that the EC calcium binding site involves residues E772-778: DADQWEE.

Formation of the human fibrinogen subclass fib420: disulfide bonds and glycosylation in its unique (alphaE chain) domains.

COS cell transfection has been used to monitor the assembly and secretion of fibrinogen molecules, both those of the subclass containing the novel alphaE chain and those of the more abundant subclass whose alpha chains lack alphaE's globular C-terminus. That region, referred to as the alphaEC domain, is closely related to the ends of beta and gamma chains of fibrinogen (betaC and gammaC). Transfection of COS cells with alphaE, beta, and gamma cDNAs alone results in secretion of the symmetrical molecule (alphaEbetagamma)2, also known as Fib420. Cotransfection with cDNA for the shorter alpha chain yielded secretion of both (alphabetagamma)2 and (alphaEbetagamma)2 but no mixed molecules of the structure alphaalphaE(betagamma)2. Exploiting the COS cells' fidelity with regard to Fib420 production, identification was made of the highly conserved Asn667 as the sole site of N-linked glycosylation in the alphaE chain. No evidence from Cys --> Ser replacements was found for interchain disulfide bridges involving the four cysteines of the alphaEC domain. However, for fibrinogen secretion, the alphaE, beta, and gamma subunits do exhibit different requirements for integrity of the two intradomain disulfide bridges located at homologous positions in their respective C-termini, indicating dissimilar structural roles in the process of fibrinogen assembly.

Assembly and secretion of fibrinogen. Involvement of amino-terminal domains in dimer formation.

Fibrinogen is a dimer with each half-molecule composed of three different chains (A alpha, B beta, gamma). Previous studies showed that amino-terminal disulfide bonds, as well as the disulfide rings that flank the "coiled-coil" region, are necessary for chain assembly and secretion (Zhang, J.Z., and Redman, C.M. (1994) J. Biol. Chem. 269, 652-658). We now determine whether other amino-terminal domains are involved in linking the half-molecules. Fibrinogen chains, with deletions at the amino terminus, were co-expressed in COS cells together with normal fibrinogen chains. Elimination of the first 8 amino acids of the B beta chain did not affect dimer assembly, but deletion of amino acid residues 9-72 had a small inhibitory effect on dimer formation. Deletion of the first 72 amino acids of the B beta chain further inhibited dimer formation and resulted in nearly equal amounts of half-molecule and dimeric fibrinogen being formed and secreted. Deletion of the first 80 residues, which includes the cysteine residues that form the amino-terminal disulfide ring, completely eliminated dimer formation, and only half-molecules were secreted. By contrast deletion of the fist 41 amino acid residues of the A alpha chain or the first 15 residues of the gamma chain, which correspond to B beta delta 1-72, did not affect chain assembly and secretion. However, co-expression of both A alpha delta 1-41 and gamma delta 1-15 with normal B beta, inhibited dimer formation. Taken together, these results indicate that in addition to disulfide bonds, noncovalent interactions of other amino-terminal amino acid residues in the three fibrinogen chains also participate in dimer formation.

Secretion of biologically active recombinant fibrinogen by yeast.

Fibrinogen (340 kDa) is a plasma protein that plays an important role in the final stages of blood clotting. Human fibrinogen is a dimer with each half-molecule composed of three different polypeptides (A alpha, 67 kDa; B beta, 57 kDa; gamma, 47 kDa). To understand the mechanism of fibrinogen chain assembly and secretion and to obtain a system capable of producing substantial amounts of fibrinogen for structure-function studies, we developed a recombinant system capable of secreting fibrinogen. An expression vector (pYES2) was constructed with individual fibrinogen chain cDNAs under the control of a Gal-1 promoter fused with mating factor F alpha 1 prepro secretion signal (SS) cascade. In addition, other constructs were prepared with combinations of cDNAs encoding two chains or all three chains in tandem. Each chain was under the control of the Gal-1 promoter. These constructs were used to transform Saccharomyces cerevisiae (INVSC1; Mat alpha his3-delta 1 leu2 trp1-289 ura3-52) in selective media. Single colonies from transformed yeast cells were grown in synthetic media with 4% raffinose to a density of 1 x 10(8) cells/ml and induced with 2% galactose for 16 h. Yeast cells expressing all three chains contained fibrinogen precursors and nascent fibrinogen and secreted about 30 micrograms/ml of fibrinogen into the culture medium. The B beta and gamma chains, but not A alpha, were glycosylated. Glycosylation of B beta and gamma chains was inhibited by treatment of transformed yeast cells with tunicamycin. Intracellular B beta and gamma chains, but not the A alpha chains in secreted fibrinogen, were cleaved by endoglycosidase H. Carbohydrate analysis indicated that secreted recombinant fibrinogen contained N-linked asialo-galactosylated biantennary oligosaccharide. Recombinant fibrinogen yielded the characteristic plasmin digestion products, fragments D and E, that were immunologically indistinct from the same fragments obtained from plasma fibrinogen. The recombinant fibrinogen was shown to be biologically active in that it could form a thrombin-induced clot, which, in the presence of factor XIIIa, could undergo gamma chain dimerization and A alpha chain polymer formation.

Topology of Kell blood group protein and the expression of multiple antigens by transfected cells.

Kell is one of the major blood group systems in human red blood cells (RBCs). The Kell antigens are carried on a 731 amino acid glycoprotein that is thought to span the erythrocyte membrane once. Rabbit antibodies to three synthetic peptides, derived from different parts of the Kell protein, were used to determine the topology of Kell protein on the RBC. Antibodies to a C-terminal peptide and to a peptide derived from amino acid residues 410 to 439 reacted with RBCs treated with 0.2 mol/L dithiothreitol. An antibody to the N-terminal peptide reacted with inside-out RBC vesicles but not with right-side-out vesicles nor with intact RBCs, showing that Kell is a type II membrane protein and that the extracellular portion of the protein is folded by disulfide bonds. By transfection, Kell protein was expressed on the cell surface of surrogate cells, and the transfected cells expressed similar antigenic properties as native RBCs. Kell protein was expressed in COS-1 and K562 cells and in Sf9 cells infected by the Baculovirus system. Transfected K562 cells expressed several high-incidence antigens but not the low-incidence antigen K1.

Role of interchain disulfide bonds on the assembly and secretion of human fibrinogen.

The cysteines involved in joining the 2 half-molecules of fibrinogen and also those located on either side of the alpha-helical coiled-coil region, were substituted, by site-directed mutagenesis, with serine. Fibrinogen assembly and secretion were determined in transiently transfected COS cells. Our studies indicate that in order to assemble the 2 half-molecules into a dimer, it is not sufficient to only have the disulfide linkages which keep the 2 half-molecules intact. The disulfide rings which flank the coil-coiled region also play important roles in dimer assembly. Intact interchain disulfide linkages at the NH2-terminal end of the coiled-coil region are essential for assembly of the 2 half-molecules. Disruption of these disulfide rings leads to the formation and secretion of half-molecules. Disruption of the interchain disulfide rings at the COOH-terminal end of the coiled-coil region allows dimer formation, but the 6-chain molecule which is assembled is not secreted. Disruption of both disulfide rings at either end of the coiled-coil region disallows assembly of half-molecules and of dimeric fibrinogen.

Symmetrical disulfide bonds are not necessary for assembly and secretion of human fibrinogen.

Human fibrinogen is a dimer, and each half-molecule is composed of three different polypeptides (A alpha, B beta, and gamma). The two half-molecules are joined together at the amino-terminal ends by three symmetrical disulfide bonds between adjacent A alpha chains, at position 28, and between adjacent gamma chains at gamma 8 and 9. To determine the role of these disulfide bonds in fibrinogen assembly and secretion, site-directed mutagenesis was used to change cysteines of A alpha 28 and gamma 8 and 9 to serine, the mutant chains were coexpressed with normal fibrinogen chains in COS cells, and assembly and secretion of fibrinogen were measured. Elimination of the symmetrical disulfide bonds did not affect assembly of the chains and dimeric fibrinogen was secreted. Analysis by plasmin digestion indicated that the secreted mutant fibrinogens have a similar structure to normal fibrinogen. Our results indicate that other domains of fibrinogen participate in dimer formation and that the three symmetrical disulfide bonds are not crucial for assembly and secretion of fibrinogen.

The human Kell blood group gene maps to chromosome 7q33 and its expression is restricted to erythroid cells.

The Kell blood group is one of the major antigenic systems in human red blood cells. To determine the location of the Kell gene on human chromosomes, panels containing genomic DNA of human-hamster somatic cell hybrids were hybridized with radiolabeled cDNA probe specific for the Kell locus. Only the samples containing DNA from chromosome 7 gave positive hybridization signals. In situ hybridization analysis, using genomic clones isolated with the cDNA, localized the KEL gene to 7q33. Northern blot analysis of poly(A)+ RNA from human brain, kidney, lung, fetal and adult liver, and bone marrow showed that Kell transcripts were only present in fetal liver and bone marrow. This indicates that the Kell protein, which carries the Kell antigens, may only be expressed in erythroid tissues.

Carboxy-terminal-extended variant of the human fibrinogen alpha subunit: a novel exon conferring marked homology to beta and gamma subunits.

Similarities between the N-terminal regions of the three subunits of the clotting protein fibrinogen--(alpha beta gamma)2--suggest that they evolved from a common progenitor. However, to date no human alpha chain has been found with the strong C-terminal homology shared by the beta and gamma chains. Here we examine the natural product of a novel fibrinogen alpha chain transcript bearing a separate open reading frame that supplies the missing C-terminal homology to the other chains. Additional splicing leads to the use of this extra sequence as a sixth exon elongating the alpha chain by 35%. Since the extended alpha chain (alpha E) is assembled into fibrinogen molecules and its synthesis is enhanced by interleukin-6, it suggests participation in both the acute phase response and normal physiology.

Molecular cloning and primary structure of Kell blood group protein.

The Kell blood group is a major antigenic system in human erythrocytes. Kell antigens reside on a 93-kDa membrane glycoprotein that is surface-exposed and associated with the underlying cytoskeleton. We isolated tryptic peptides and, based on the amino acid sequence of one of the peptides and by using the PCR, prepared a specific oligonucleotide to screen a lambda gt10 human bone-marrow cDNA library. Four clones were isolated, one containing cDNA with an open reading frame for an 83-kDa protein. All known Kell amino acid sequences were present in the deduced sequence; moreover, rabbit antibody to a 30-amino acid peptide, prepared from this sequence, reacted on an immunoblot with authentic Kell protein. The Kell cDNA sequence predicts a 732-amino acid protein. Hydropathy analysis indicates a single membrane-spanning region, suggesting that Kell protein is oriented with 47 of its N-terminal amino acids in the cell cytoplasm, and a 665-amino acid segment, which contains six possible N-glycosylation sites, is located extracellularly. Computer-based search showed that Kell has structural and sequence homology to a family of zinc metalloglycoproteins with neutral endopeptidase activity.

Regulation of fibrinogen assembly. Transfection of Hep G2 cells with B beta cDNA specifically enhances synthesis of the three component chains of fibrinogen.

Previous studies indicated that synthesis of B beta chain may be a rate-limiting factor in the production of human fibrinogen since Hep G2 cells contain surplus pools of A alpha and gamma but not of B beta chains, and fibrinogen assembly commences by the addition of preformed A alpha and gamma chains to nascent B beta chains attached to polysomes. To test whether B beta chain synthesis is rate limiting Hep G2 cells were transfected with B beta cDNA, and its effect on fibrinogen synthesis and secretion was measured. Two sets of stable B beta cDNA-transfected Hep G2 cells were prepared, and both cell lines synthesized 3-fold more B beta chains than control cells. The B beta-transfected cells also synthesized and secreted increased amounts of fibrinogen. Transfection with B beta cDNA not only increased the synthesis of B beta chain but also increased the rate of synthesis of the other two component chains of fibrinogen and maintained surplus intracellular pools of A alpha and gamma chains. Transfection with B beta cDNA did not affect the synthesis of albumin, transferrin, or anti-chymotrypsin and had a small inhibitory effect on the synthesis of C-reactive protein. Taken together these studies demonstrate that increased B beta chain synthesis specifically causes increased production of the other two component chains of fibrinogen and that unequal and surplus amounts of A alpha and gamma chains are maintained intracellularly.

Localization of the McLeod locus (XK) within Xp21 by deletion analysis.

The McLeod phenotype is an X-linked, recessive disorder in which the red blood cells demonstrate acanthocytic morphology and weakened antigenicity in the Kell blood group system. The phenotype is associated with a reduction of in vivo red cell survival, but the permanent hemolytic state is usually compensated by erythropoietic hyperplasia. The McLeod phenotype is accompanied by either a subclinical myopathy and elevated creatine kinase (CK) or X-linked chronic granulomatous disease (CGD). Seven males with the McLeod red-blood-cell phenotype and associated myopathy but not CGD, one male with the McLeod phenotype associated with CGD, and two males known to possess large deletions of the Duchenne muscular dystrophy (DMD) locus were studied. DNA isolated from each patient was screened for the presence or absence of various cloned sequences located in the Xp21 region of the human X chromosome. Two of the seven males who have only the McLeod phenotype and are cousins exhibit deletions for four Xp21 cloned fragments but are not deleted for any portion of either the CGD or the DMD loci. Comparison of the cloned segments absent from these two McLeod cousins with those absent from the two DMD boys and the CGD/McLeod patient leads to the submapping of various cloned DNA segments within the Xp21 region. The results place the locus for the McLeod phenotype within a 500-kb interval distal from the CGD locus toward the DMD locus.

Regulation of apo-A-I processing in cultured hepatocytes.

Apo-A-I, the major protein component of high density lipoproteins, appears intracellularly as an intermediate precursor (pro-apo-A-I) with a hexapeptide extension (RHFWQQ) at its amino terminus. Proteolytic processing of pro-apo-A-I to apo-A-I has been shown to occur extracellularly in cell and organ cultures from rat and human tissues. Recently, however, intracellular conversion has been detected in chickens. To determine what distinguishes and regulates these two processing methods, the proteolytic processing and secretion of apo-A-I was studied by metabolic labeling in chick hepatocytes and in Hep-G2 cells (derived from a human hepatocellular carcinoma). The proportions of intracellular and secreted pro-apo-A-I and apo-A-I were measured by sequencing NH2-terminal portions of the proteins and determining the location of radio-labeled amino acids. Chick hepatocytes cultured in the absence of hormones or fetal bovine serum secreted primarily processed apo-A-I (83%). In the presence of serum these cells secreted only pro-apo-A-I, whereas incubation with a combination of hormones (insulin, triiodothyronine, dexamethasone) resulted in secretion of a nearly equal mixture of the pro- and processed forms of the protein. In contrast, Hep-G2 cells, maintained in the absence of serum, secreted only pro-apo-A-I; when grown in the presence of serum these cells secreted a mixture of pro- and processed apo-A-I. Under conditions in which chick hepatocytes and Hep-G2 cells secreted both forms of the protein, a mixture of pro- and processed apo-A-I was also found intracellularly; when only the pro-form was secreted, the cells likewise contained only pro-apo-A-I. Under all the above conditions, the secreted apo-A-I exhibited similar isoform patterns in two-dimensional gel electrophoresis. These data show that both chick hepatocytes and human hepatoma cells are capable of intracellularly processing pro-apo-A-I to apo-A-I, and that the extent of intracellular processing is controlled by the cell's hormonal environment.

Biosynthesis of spectrin and its assembly into the cytoskeletal system of Friend erythroleukemia cells.

Friend erythroleukemia cells, grown in the presence of dimethyl sulfoxide for 3 d, synthesize unequal amounts of the two chains (alpha and beta) of spectrin with approximately 15-30% more beta than alpha spectrin. When cells were ruptured by nitrogen cavitation, nascent alpha and beta spectrin were found to be associated with a membranous cell fraction and were not detected in the soluble cytoplasmic cell fraction. Nascent membrane-bound spectrin appeared not to be protected by membranes, since it was susceptible to trypsin degradation in the absence of detergent. On fractionation of cells with 1% Triton X-100, more (1.75-fold) nascent spectrin was found in the Triton-soluble fraction than in the Triton-insoluble fraction (cytoskeleton). In the Triton-soluble fraction, there was 55% more nascent beta spectrin than alpha spectrin, while the cytoskeleton contained nearly equal amounts of alpha and beta spectrin. Cells were pulse-labeled with L-[35S]methionine for 2 min and chase incubated for varying periods of time from 15 to 90 min with nonradioactive L-methionine. Radioactive spectrin accumulated in the Triton-soluble fraction for the first 15 min of chase incubation and then dropped by 25% in the next hour. By contrast, the amount of radioactive spectrin in the Triton-insoluble fraction rose gradually for 1 h of the chase period. This indicates that, in Friend erythroleukemia cells, a pool of membrane-bound spectrin containing an excess of the beta polypeptide is used to form the cytoskeletal system which is composed of equal molar amounts of alpha and beta spectrin. The location of spectrin was determined by immunoelectron microscopy. Small amounts of spectrin were detected in cells not treated with dimethyl sulfoxide and in these cells it was located on the surface membrane and within the cytoplasm. On treatment with dimethyl sulfoxide, complex vacuolar structures containing viruses appeared in the cells. In cells treated with dimethyl sulfoxide for 3 d 30% of the spectrin was near the outer membrane and 25% was associated with vacuolar structures, whereas in cells treated for 5 and 7 d the majority of spectrin (57-61%) was located in the vacuolar areas.

Tyrosine sulfation of proteins from the human hepatoma cell line HepG2.

[35S]Sulfate labeling of the human hepatoma cell line HepG2 showed it to contain many sulfated proteins of diverse molecular weight range. The isolation of tyrosine O-sulfate indicated the supernatant fraction to contain a 5- to 7-fold higher level than the cellular fraction at the end of a 24-hr incubation. The proteins in the supernatant fraction were immunoprecipitated and examined for sulfation. Of 15 proteins tested, 7 were found to be sulfated as indicated by [35S]sulfate incorporation into proteins separated by NaDodSO4/PAGE and detected by autoradiography. The 35S-labeled bands were excised from the dried gel and subjected to extensive Pronase hydrolysis and the hydrolysates were analyzed for tyrosine [35S]sulfate by a two-dimensional procedure combining high-voltage electrophoresis and thin-layer chromatography [Liu, M. C. & Lipmann, F. (1984) Proc. Natl. Acad. Sci. USA 81, 3695-3698]. Of the sulfated proteins, three--fibrinogen, alpha-fetoprotein, and fibronectin--were found to contain tyrosine O-sulfate. The simultaneous presence of carbohydrate-bound sulfate, however, could not be exactly determined, but the other four [35S]sulfate-containing proteins--alpha 1-antitrypsin, alpha 1-antichymotrypsin, alpha 2-macroglobulin, and transferrin--did not reveal any tyrosine O-sulfate and might be sulfated on their carbohydrate moieties.