Transforming growth factor alpha in arterioles: cell surface processing of its precursor by elastases.

Analysis of the transforming growth factor alpha (TGF alpha) cDNA predicts that the mature TGF alpha polypeptide is cleaved from the extracellular domain of its precursor, which is an integral membrane protein. Furthermore, the cleavage sites for the release of this mitogen are compatible with the participation of an elastaselike protease. We have immunohistochemically localized TGF alpha to the vascular smooth muscle cells in the arterioles. To investigate whether polymorphonuclear (PMN) leukocytic elastase, a blood-borne protease, could process the cell surface TGF alpha, NR6 cells were transfected with the rat TGF alpha cDNA. The cDNA encoded the entire open reading frame, and its expression was under the control of the mouse metallothionein I promoter. A cloned transfectant, termed 1B2, synthesized the TGF alpha precursor in a zinc-inducible manner, and the precursor was localized to the cell surface. Western blot (immunoblot) analysis indicated that treatment of the zinc-induced 1B2 cells with either PMN leukocytic or pancreatic elastase resulted in the release of the mature TGF alpha polypeptide. The released TGF alpha was bioactive, as it was capable of both competing with epidermal growth factor for binding to its receptor and stimulating [3H]thymidine incorporation in the mitogenic assay. Formaldehyde fixation of the 1B2 cells eliminated basal release of TGF alpha but allowed normal processing by both PMN leukocytic and pancreatic elastase to occur. However, human cathepsin G, bovine pancreatic alpha 1-chymotrypsin, collagenase, trypsin, subtilisin, and plasmin failed to release any detectable fragments of the TGF alpha precursor from the fixed cells. The location of TGF alpha in the arterioles and ability of PMN leukocytic elastase to process the membrane-bound TGF alpha precursor suggests a novel role for this elastase at the wound site.     

Endogenous meltrin alpha is ubiquitously expressed and associated with the plasma membrane but exogenous meltrin alpha is retained in the endoplasmic reticulum

Meltrin alpha (ADAM12) is a member of the ADAM (MDC) protein family characterized by the presence of metalloprotease and disintegrin domains. ADAM proteins contain single transmembrane domains, and the processed mature proteins are postulated to span the plasma membrane. It has been reported that transfection of a truncated meltrin alpha cDNA lacking the prodomain and metalloprotease domain promotes skeletal muscle cell fusion. We show here that meltrin alpha was constitutively expressed in both undifferentiated and differentiated C2 skeletal muscle cells and also in fibroblasts. Both its precursor and processed mature forms were present in these cells. Thus, meltrin alpha may play general roles in addition to its roles in myogenesis. Since endogenous meltrin alpha cannot be detected by immunofluorescence microscopy, we examined the location of the exogenously expressed protein by transfection. Unexpectedly, the exogenously expressed meltrin alpha was located to a network structure of the endoplasmic reticulum (ER) but not to the plasma membrane. Cell fractionation revealed that the intrinsic mature protein was associated with the plasma membrane. However, the exogenously expressed protein remained unprocessed. These results seem to imply that the exogenously expressed meltrin alpha is not translocated from the ER to the trans-Golgi network, where a processing enzyme resides, and that it is consequently not converted to the mature form. Thus, the transfected meltrin alpha is unlikely to exert its physiological functions. Conversely, the ER may serve as a reservoir of the latent form of intrinsic meltrin alpha.     

Tumorigenic transformation of NIH 3T3 cells by the autocrine synthesis of transforming growth factor alpha

A variety of cancer cells overexpress transforming growth factor alpha (TGF alpha), a mitogenic peptide. A cDNA sequence coding for the full-length human TGF alpha precursor protein was subcloned into a retroviral expression vector and introduced into clone 7 NIH 3T3 cells, which have low numbers of endogenous epidermal growth factor receptors (EGFRs). The autocrine synthesis of TGF alpha by these cells resulted in their focal transformation. In contrast, control NIH 3T3 cells treated in a paracrine manner with exogenous, saturating concentrations of the mature form of TGF alpha, though stimulated to divide, remained morphologically untransformed. The addition of saturating quantities of soluble, mature TGF alpha to NIH 3T3 cells expressing the transferred TGF alpha gene actually suppressed their growth and focal transformation. The transformation induced by the TGF alpha gene remained an EGFR-dependent process, since the degree of transformation was correlated with EGFR expression in NIH 3T3 cells and since NR6 cells, which are Swiss 3T3 cells devoid of endogenous EGFRs, were transformed by the TGF alpha vector only when exogenous EGFR genes were also introduced. When inoculated into nude mice, the TGF alpha-expressing cells rapidly gave rise to tumors that grew progressively, whereas control cells did not form tumors. We conclude that in certain circumstances autocrine TGF alpha can be more oncogenic than paracrine and that paracrine TGF alpha can suppress this effect.     

Human lysosomal acid phosphatase is transported as a transmembrane protein to lysosomes in transfected baby hamster kidney cells.

BHK cells transfected with human lysosomal acid phosphatase (LAP) cDNA (CT29) expressed 70-fold higher enzyme activities of acid phosphatase than non-transfected BHK cells. The CT29-LAP was synthesized in BHK cells as a heterogeneously glycosylated precursor that was tightly membrane associated. Transfer to the trans-Golgi was associated with a small increase in size (approximately 7 kd) and partial processing of the oligosaccharides to complex type structures. CT29-LAP was transferred into lysosomes as shown by subcellular fractionation, immunofluorescence and immunoelectron microscopy. Lack of mannose-6-phosphate residues suggested that transport does not involve mannose-6-phosphate receptors. Part of the membrane-associated CT29-LAP was processed to a soluble form. The mechanism that converts CT29-LAP into a soluble form was sensitive to NH4Cl, and reduced the size of the polypeptide by 7 kd. In vitro translation of CT29-derived cRNA in the presence of microsomal membranes yielded a CT29-LAP precursor that is protected from proteinase K except for a small peptide of approximately 2 kd. In combination with the sequence data available for LAP, these observations suggest that CT29-LAP is synthesized and transported to lysosomes as a transmembrane protein. In the lysosomes, CT29-LAP is released from the membrane by proteolytic cleavage, which removes a C-terminal peptide including the transmembrane domain and the cytosolic tail of 18 amino acids.     

A novel low molecular weight ribonucleic acid (RNA) related to transforming growth factor alpha messenger RNA

Human transforming growth factor alpha (TGF alpha) is coded for by an mRNA of about 4800 nucleotides. The cDNA sequence demonstrates that the 50 amino acid TGF alpha is embedded in a larger 160 amino acid precursor protein. We report here that in addition to the 4800 nucleotide TGF alpha mRNA, there is a novel second RNA species of about 350 nucleotides that hybridizes to a human TGF alpha cDNA probe. This small RNA species has been found in the RNA of several human tumor cells including HT1080, A549, A431, A2058, and A673. We have demonstrated an inverse relationship between the amounts of the 4800 nucleotide TGF alpha mRNA and the 350 nucleotide novel RNA in these human cells. Restriction enzyme cleavage of a human TGF alpha cDNA probe into three separate domains consisting of a processed coding region and 5'- and 3'-preprocessed coding and untranslated regions showed that only the 3'-untranslated region hybridized to the 350 nucleotide RNA. Using sense and anti-sense single-stranded 3'-untranslated region probes, we determined that the 350 nucleotide RNA band may be composed of multiple species of RNA which are related to the anti-sense DNA strand that is opposite to the strand that codes for the 4800 nucleotide TGF alpha mRNA.     

Functional integrins from normal and glycosylation-deficient baby hamster kidney cells. Terminal processing of asparagine-linked oligosaccharides is not correlated with fibronectin-binding activity.

We have examined the properties of the alpha 5 beta 1 integrin of baby hamster kidney (BHK) cells, a ricin-resistant variant Ric14 lacking N-acetylglucosaminyl transferase I, and hence unable to complete assembly of hybrid- or complex-type N-glycans, and BHK cells treated with 1-deoxymannojirimycin (dMM), an inhibitor of Golgi mannosidases involved in the initial processing of N-glycan precursors. Comparable amounts of alpha 5 beta 1 integrin were isolated from these cells by chromatography of detergent extracts on a fibronectin cell-binding fragment affinity column and elution with EDTA. The alpha 5 beta 1 integrin obtained from normal BHK cells by fibronectin affinity chromatography contained mainly endoglycosidase H-resistant oligosaccharides, whereas in RicR14 cells or dMM-treated BHK cells these were entirely endoglycosidase H-sensitive. Analysis of lactoperoxidase labeled or long term biosynthetically 35S-labeled proteins from cultures of normal or glycosylation deficient cells showed similar steady state levels of alpha 5 beta 1 integrin and expression at the cell surface. Pulse-chase experiments in normal BHK cells showed rapid conversion of the alpha 5 subunit into a mature form containing oligosaccharides resistant to endoglycosidase H and slower maturation of a precursor beta 1 subunit, as in other cell types. In Ric14 cells the precursor beta 1 subunit was found to carry glycans larger than the fully processed Man5GlcNAc2 glycan of the mature subunit, indicating that the bulk precursor pool had not been translocated into the cis-Golgi compartment containing mannosidase I. We conclude that in BHK cells terminal oligosaccharide processing of alpha 5 beta 1 integrin subunits is not required for dimer formation, surface expression, and fibronectin binding, and that expression of the glycosylation defect of Ric14 cells on the alpha 5 beta 1 integrin does not account for the reduced adhesiveness of these cells on fibronectin compared with normal and dMM-treated BHK cells.     

Secretion of soluble functional insulin receptors by transfected NIH3T3 cells

In order to determine whether the human insulin receptor ectodomain can be expressed as a functional protein, the coding regions for the transmembrane and cytoplasmic domain of a full-length human insulin receptor cDNA were deleted by site-directed mutagenesis, and the resultant construct was inserted into a bovine papilloma virus vector under the control of the mouse metallothionein promoter. After transfection of mouse NIH3T3 cells, a cell line secreting an insulin binding protein was isolated. The insulin binding alpha subunit had an Mr of 138,000 and a beta subunit of Mr 48,000 (compared to 147,000 and 105,000 for the full-length human insulin receptor expressed in NIH3T3 cells). This difference in size of the alpha subunit was due to a difference in glycosylation as N-glycanase digestion reduced the apparent size of the alpha subunits of secreted and normal membrane-bound receptors to identical values. The secreted receptor formed disulfide-linked heterotetrameric structures with an Mr of 280,000. It was synthesized as an Mr 160,000 precursor which was cleaved into mature subunits with a t1/2 of 3 h. Increasing expression of the cDNA by induction with sodium butyrate lead to the appearance of an Mr 180,000 protein in the medium as well as the mature alpha and beta subunits. A Scatchard plot of insulin binding to the secreted receptor was curvilinear with a Kd of 7 X 10(-10) M for the high affinity sites and 10(-7) M for the low affinity site (compared to Kd values of 1.1 X 10(-9) M and 10(-7) M, respectively, for human insulin receptors expressed in these cells.     

Characterization of high molecular weight transforming growth factor alpha produced by rat hepatocellular carcinoma cells

In addition to the mature 50 amino acid transforming growth factor alpha (TGF alpha), some transformed cells appear to produce multiple higher molecular weight forms. The structure and derivation of most of these larger soluble TGF alpha species remain to be established. We previously reported that a chemically induced rat hepatocellular carcinoma cell line, JM1, secreted acid-stable proteins which bind to epidermal growth factor receptors and stimulate DNA synthesis in primary cultures of normal adult rat hepatocytes. Purification and characterization of these hepatoma-derived growth factors have indicated their relationship to TGF alpha. Two EGF-competing activities of apparent Mr 30K and 10K were separated by gel filtration of concentrated JM1-conditioned medium and further purified by ion-exchange chromatography and reverse-phase HPLC. Both growth factors were detected by a radioimmunoassay specific for TGF alpha. Western blotting with antibodies to the 50 amino acid TGF alpha revealed that the lower molecular weight factor comigrated with the synthetic 6-kDa rat TGF alpha. The higher molecular weight TGF alpha appeared on immunoblots as a diffuse band of 18-21 kDa, which converted to the mature 6-kDa form upon digestion with elastase, confirming a precursor-product relationship. However, the 18-21-kDa proteins did not react with antibodies directed against the carboxy-terminal cytoplasmic segment of the transmembrane TGF alpha precursor. Enzymatic deglycosylation of the 18-21-kDa TGF alpha species by sequential removal of sialic acids and O- and N-linked carbohydrate reduced the molecular weight to 11K. The size and soluble nature of this polypeptide suggest that it represents the extracellular domain of the transmembrane TGF alpha precursor.(ABSTRACT TRUNCATED AT 250 WORDS)     

Cloning of a growth arrest-specific and transforming growth factor beta-regulated gene, TI 1, from an epithelial cell line.

By cDNA cloning and differential screening, five genes that are regulated by transforming growth factor beta (TGF beta) in mink lung epithelial cells were identified. A novel membrane protein gene, TI 1, was identified which was downregulated by TGF beta and serum in quiescent cells. In actively growing cells, the TI 1 gene is rapidly and transiently induced by TGF beta, and it is overexpressed in the presence of protein synthesis inhibitors. It appears to be related to a family of transmembrane glycoproteins that are expressed on lymphocytes and tumor cells. The four other genes were all induced by TGF beta and correspond to the genes of collagen alpha type I, fibronectin, plasminogen activator inhibitor 1, and the monocyte chemotactic cell-activating factor (JE gene) previously shown to be TGF beta regulated.     

Stromelysin, a connective tissue-degrading metalloendopeptidase secreted by stimulated rabbit synovial fibroblasts in parallel with collagenase. Biosynthesis, isolation, characterization, and substrates

Rabbit synovial fibroblasts induced to undergo a specific switch in gene expression by agents that alter cell morphology secreted the neutral proteinase precursor procollagenase (apparent Mr of 53,000 and 57,000). A major Mr = 51,000 polypeptide that was always induced coordinately with procollagenase has now been identified as the proenzyme form of a metal-dependent proteinase active at neutral pH. We have named this proteinase stromelysin. Prostromelysin and procollagenase were the most prominent [35S]methionine-labeled secreted proteins of the induced fibroblasts. By the use of casein degradation as an assay for enzyme activity, stromelysin was isolated with high yield from the conditioned culture medium of 12-O-tetradecanoylphorbol 13-acetate-treated fibroblasts and migrated as an active form of Mr = 21,000 that was immunologically identical to the proteoglycan-degrading proteinase purified from rabbit bone. Immunoglobulin G from antiserum raised to purified rabbit bone proteoglycanase immunoprecipitated the Mr = 51,000 proenzyme form from conditioned medium of induced rabbit cells and also immunoprecipitated an Mr = 55,000 polypeptide from induced human fibroblasts. When rabbit prostromelysin was activated by trypsin or 4-aminophenylmercuric acetate, the proenzyme was converted to an active form of Mr = 41,000. During the course of the purification, prostromelysin was converted to an additional activatable form of Mr = 35,000 and additional active forms of Mr = 21,000-25,000, which had related peptide maps distinct from collagenase. All of these forms were immunologically cross-reactive. Purified stromelysin degraded casein, cartilage proteoglycans, fibronectin, alpha 1-proteinase inhibitor, and immunoglobulin G2a and had limited activity on laminin, elastin, type IV collagen, and gelatin, but did not degrade type I collagen. Stromelysin was inhibited by EDTA, 1,10-phenanthroline, and the specific glycoprotein tissue inhibitor of metalloproteinases isolated from human amniotic fluid and was therefore classified as a metalloproteinase.     

Cytotoxic activities of a fusion protein comprised of TGF alpha and Pseudomonas exotoxin

A cDNA encoding transforming growth factor type alpha (TGF alpha) was fused to the 5' end of a gene encoding a modified form of Pseudomonas exotoxin A (PE), which is devoid of the cell recognition domain (domain Ia). The chimeric molecule, termed TGF alpha-PE40, was expressed in Escherichia coli and isolated from the periplasm or inclusion bodies depending on the construction expressed. TGF alpha-PE40 was found to be extremely cytotoxic to cells displaying epidermal growth factor (EGF) receptors. Comparison with a similar molecule in which TGF alpha was placed at the carboxyl end of PE40 demonstrated the importance of the position of the cell recognition element; TGF alpha-PE40 was found to be about 30-fold more cytotoxic to cells bearing EGF receptors than PE40-TGF alpha. In addition, TGF alpha-PE40 was shown to be extremely cytotoxic to a variety of cancer cell lines including liver, ovarian, and colon cancer cell lines, indicating high levels of EGF receptor expression in these cells.     

Phylogenetic conservation of arylsulfatases. cDNA cloning and expression of human arylsulfatase B

A 2.2-kilobase cDNA clone for human arylsulfatase B (ASB) and several genomic clones were isolated and sequenced. The deduced amino acid sequence of 533 amino acids contains a 41-amino acid N-terminal signal peptide and a mature polypeptide of 492 amino acid residues. Overexpression of ASB in transfected baby hamster kidney (BHK) cells resulted in up to 68-fold higher ASB activity than in untransfected BHK cells. Pulse-chase labeling showed that ASB was synthesized and secreted as a 64-kDa precursor and processed to a 47-kDa mature form in BHK cells. The 47-kDa ASB form was located in dense lysosomes. Transport of ASB to the lysosomes was accomplished in a mannose 6-phosphate receptor-dependent manner. The ASB cDNA clone hybridizes to 4.8-, 2.5-, and 1.8-kilobase species of RNA from human fibroblasts. The same pattern was observed in RNA from fibroblasts of three Maroteaux-Lamy patients who were deficient in ASB activity, as well as in RNA from fibroblasts of three patients with multiple sulfatase deficiency, in which all known sulfatases were markedly diminished. Deduced amino acid sequences of human arylsulfatase A, human ASB, human steroid sulfatase, human glucosamine-6-sulfatase, and an arylsulfatase from sea urchin showed a substantial degree of similarity suggesting that they arose from a common ancestral gene and are members of an arylsulfatase gene family.     

Estrogen receptor alpha mediated induction of the transforming growth factor alpha gene by estradiol and 4-hydroxytamoxifen in MDA-MB-231 breast cancer cells

The selective estrogen receptor modulator, 4-hydroxytamoxifen (4-OHT) is a full agonist at the transforming growth factor (TGF) alpha gene in ER negative breast cancer cells stably transfected with ER alpha cDNA (Levenson et al., Br. J. Cancer 77 (1998) 1812-1819). E(2) and 4-OHT increase TGF alpha mRNA and protein in a concentration dependent manner. The responses to E(2) and 4-OHT are blocked by the pure antiestrogen ICI 182,780, which does not induce TGF alpha. Transfected MDA-MB-231 cells contain functional ER alpha but no ER beta function was detected. Neo transfected cells that did not express ER alpha or cells stably transfected with the DNA binding domain mutant C202R/E203V which prevents gene activation did not induce TGF alpha mRNA after either E(2) or 4-OHT treatment. An examination of the time course for either 10 nM E(2) or 1 microM 4-OHT for MDA-MB-231 cells stably transfected with cDNA for ER alpha showed increases in TGF alpha mRNA within 2 or 3 h respectively. Cells pretreated with cycloheximide (1 microg/ml) showed induced TGF alpha mRNA in response to E(2) or 4-OHT but TGF alpha mRNA induction was blocked by actinomycin D (1 microg/ml). We conclude that both E(2) and 4-OHT induce TGF alpha by direct interaction of ER alpha with DNA and that ER beta is not involved in the estrogen-like response to 4-OHT in the MDA-MB-231 cells.     

Binding of transforming growth factor-beta to cell surface proteins varies with cell type

Transforming growth factor-beta (TGF beta 1 and TGF beta 2) bind to several different cell surface proteins, including a high Mr proteoglycan. We found that on primary and early passage cultures of fibroblasts, chondroblasts, and osteoblasts TGF beta 1 binds to both the high Mr proteoglycan and to lower Mr components, whereas on epithelial, endothelial, and lymphoid-derived cells TGF beta 1 only binds to the lower Mr species. With cell lines, this distinction is lost. Further analysis indicated that binding to the high Mr proteoglycan is not necessary for TGF beta 1 induced regulation of DNA, collagen and fibronectin synthesis, change in cell morphology, or reorganization of the actin cytoskeleton. We propose that the lower Mr components are the active receptors mediating these events.     

Cloning and sequence analyses of cDNAs for interferon- and virus-induced human Mx proteins reveal that they contain putative guanine nucleotide-binding sites: functional study of the corresponding gene promoter.

The human protein p78 is induced and accumulated in cells treated with type I interferon or with some viruses. It is the human homolog of the mouse Mx protein involved in resistance to influenza virus. A full-length cDNA clone encoding the human p78 protein was cloned and sequenced. It contained an open reading frame of 662 amino acids, corresponding to a polypeptide with a predicted molecular weight of 75,500, in good agreement with the Mr of 78,000 determined on sodium dodecyl sulfate gels for the purified natural p78 protein. The cloned gene was expressed in vitro and corresponded in size, pI, antigenic determinant(s), and NH2 terminus sequence to the natural p78 protein. A second cDNA was cloned which encoded a 633-amino-acid protein sharing 63% homology with human p78. This p78-related protein was translated in reticulocyte lysates where it shared an antigenic determinant(s) with p78. A putative 5' regulatory region of 83 base pairs contained within the gene promoter region upstream of the presumed p78 mRNA cap site conferred human alpha interferon (IFN-alpha) inducibility to the cat reporter gene. The p78 protein accumulated to high levels in cells treated with IFN-alpha. In contrast, the p78-related protein was not expressed at detectable levels. The rate of decay of p78 levels in diploid cells after a 24-h treatment with IFN-alpha was much slower than the rate of decay of the antiviral state against influenza A virus and vesicular stomatitis virus, suggesting that the p78 protein is probably not involved in an antiviral mechanism. Furthermore, we showed that these proteins, as well as the homologous mouse Mx protein, possess three consensus elements in proper spacing, characteristic of GTP-binding proteins.     

Activation of mammalian cyclic AMP phosphodiesterases by trypsin.

BHK fibroblasts contain two forms of cyclic AMP phosphodiesterase 3':5'-cyclic nucleotide 5'-nucleotidohydrolase EC 3.1.4.17) as analyzed by linear sucrose gradient fractionation; a 3.6-S form (peak I) and a 6.7-S form (peak II). Peak I is specific for cyclic AMP as substrate and displays Michaelis-Menten kinetics with an apparent Km of 2--3 micrometer. Peak II hydrolyzes cyclic GMP and displays anomalous kinetics for cyclic AMP hydrolysis. The activity of isolated peak II for cyclic AMP is increased by storage at 4 degrees C, treatment with trypsin, or treatment with rat brain and BHK fibroblast activator proteins. The activity of isolated peak I is unaffected by these conditions. Linear sucrose gradient fractionation demonstrates that activation of peak II by trypsin leads to the formation of a 3.6-S cyclic AMP-specific enzyme form, possibly peak I. In contrast to BHK fibroblasts (and most other mammalian tissues), rat uterus contains only one form of cyclic nucleotide phosphodiesterase on linear sucrose gradients, a 7-S form capable of hydrolyzing both cyclic AMP and cyclic GMP. Treatment of rat uterine supernatant with trypsin leads to the appearance of a 4-S, cyclic AMP-specific form with properties similar to that of BHK peak I. These data suggest that the kinetically complex, higher molecular weight cyclic nucleotide phosphodiesterases may consist of more than one catalytically active site and that multiple forms of the enzyme arise through dissociative mechanisms, possibly as a means of in vivo regulation.     

Differential processing of colony-stimulating factor 1 precursors encoded by two human cDNAs.

The biosynthesis of macrophage colony-stimulating factor 1 (CSF-1) was examined in mouse NIH-3T3 fibroblasts transfected with a retroviral vector expressing the 554-amino-acid product of a human 4-kilobase (kb) CSF-1 cDNA. Similar to results previously obtained with a 1.6-kb human cDNA that codes for a 256-amino-acid CSF-1 precursor, the results of the present study showed that NIH-3T3 cells expressing the product of the 4-kb clone produced biologically active human CSF-1 and were transformed by an autocrine mechanism when cotransfected with a vector containing a human c-fms (CSF-1 receptor) cDNA. The 4-kb CSF-1 cDNA product was synthesized as an integral transmembrane glycoprotein that was assembled into disulfide-linked dimers and rapidly underwent proteolytic cleavage to generate a soluble growth factor. Although the smaller CSF-1 precursor specified by the 1.6-kb human cDNA was stably expressed as a membrane-bound glycoprotein at the cell surface and was slowly cleaved to release the extracellular growth factor, the cell-associated product of the 4-kb clone was efficiently processed to the secreted form and was not detected on the plasma membrane. Digestion with glycosidic enzymes indicated that soluble CSF-1 encoded by the 4-kb cDNA contained both asparagine(N)-linked and O-linked carbohydrate chains, whereas the product of the 1.6-kb clone had only N-linked oligosaccharides. Removal of the carbohydrate indicated that the polypeptide chain of the secreted 4-kb cDNA product was longer than that of the corresponding form encoded by the smaller clone. These differences in posttranslational processing may reflect diverse physiological roles for the products of the two CSF-1 precursors in vivo.     

Processing of the precursor to a chloroplast ribosomal protein made in the cytosol occurs in two steps, one of which depends on a protein made in the chloroplast.

In pulse-chase experiments in which log-phase cells of Chlamydomonas reinhardtii were labeled in vivo for 5 min with H2(35)SO4, fluorographs of immunoprecipitates from whole cell extracts revealed that chloroplast ribosomal proteins L-2, L-6, L-21, and L-29, which are made in the cytosol and imported, appeared in their mature forms. However, in the case of chloroplast ribosomal protein L-18, which is also made in the cytoplasm and imported, a prominent precursor with an apparent molecular weight of 17,000 was found at the end of a 5-min pulse. This precursor was processed to its mature size (apparent molecular weight of 15,500) within the first 5 min of the subsequent chase. As determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the precursor to L-18 formed in vivo was 1.5 kilodaltons smaller than the primary product detected in translations of Chlamydomonas polyadenylated RNA in vitro. Upon a 10-min incubation with a postribosomal supernatant from Chlamydomonas, the 18,500-dalton precursor detected in vitro could be partially converted into a polypeptide that comigrated with the 17,000-dalton precursor detected in extracts of cells labeled in vivo. Under conditions in which the total amounts of chloroplast proteins had been reduced and cells were made to synthesize ribosomes rapidly, the apparent half-life of the 17,000-dalton precursor was extended over that seen in log-phase cells. When chloroplast protein synthesis was inhibited with lincomycin for 3 h before labeling under these conditions, the 17,000-dalton L-18 precursor but not the mature form was found, and the precursor was slowly degraded during a 60-min chase. When cells were placed in the dark for 3 h before labeling, processing of this precursor to the mature form appeared unaffected, but the chloroplast-synthesized ribosomal protein L-26 was detected, indicating that chloroplast protein synthesis was still occurring. We interpret these results to indicate that the maturation of protein L-18 in vivo involves at least two processing steps, one of which depends on a protein made on chloroplast ribosomes.