Transgenic expression of epidermal growth factor and keratinocyte growth factor in beta-cells results in substantial morphological changes

The upregulation of a limited number of growth factors in our interferon-gamma transgenic model for regeneration within the pancreas lead us to propose that these factors are important during pancreatic regeneration. In this study, we have assessed the influence of two growth factors within the pancreas, epidermal growth factor (EGF) and keratinocyte growth factor (KGF), by ectopically expressing these proteins under the control of the human insulin promoter in transgenic mice. This beta-cell-targeted expression of either EGF or KGF resulted in significant morphological changes, including cellular proliferation and disorganized islet growth. Intercrossing the individual Ins-EGF and Ins-KGF transgenic mice resulted in more profound changes in pancreatic morphology including proliferation of pancreatic cells and extensive intra-islet fibrosis. Insulin-producing beta-cells were found in some of the ducts of older Ins-EGF and Ins-EGFxKGF transgenic mice, and amylase-producing cells were observed within the islet structures of the double transgenic mice. These data suggest that both EGF and KGF are capable of affecting pancreatic differentiation and growth, and that co-expression of these molecules in islets has a more substantial impact on the pancreas than does expression of either growth factor alone.

Crystal structure of the kinase domain of human vascular endothelial growth factor receptor 2: a key enzyme in angiogenesis

BACKGROUND

Angiogenesis is involved in tumor growth, macular degeneration, retinopathy and other diseases. Vascular endothelial growth factor (VEGF) stimulates angiogenesis by binding to specific receptors (VEGFRs) on the surface of vascular endothelial cells. VEGFRs are receptor tyrosine kinases that, like the platelet-derived growth factor receptors (PDGFRs), contain a large insert within the kinase domain.

RESULTS

We report here the generation, kinetic characterization, and 2.4 A crystal structure of the catalytic kinase domain of VEGF receptor 2 (VEGFR2). This protein construct, which lacks 50 central residues of the 68-residue kinase insert domain (KID), has comparable kinase activity to constructs containing the entire KID. The crystal structure, determined in an unliganded phosphorylated state, reveals an overall fold and catalytic residue positions similar to those observed in other tyrosine-kinase structures. The kinase activation loop, autophosphorylated on Y1059 prior to crystallization, is mostly disordered; however, a portion of it occupies a position inhibitory to substrate binding. The ends of the KID form a beta-like structure, not observed in other known tyrosine kinase structures, that packs near to the kinase C terminus.

CONCLUSIONS

The majority of the VEGFR2 KID residues are not necessary for kinase activity. The unique structure observed for the ends of the KID may also occur in other PDGFR family members and may serve to properly orient the KID for signal transduction. This VEGFR2 kinase structure provides a target for design of selective anti-angiogenic therapeutic agents.

Characterization and kinetic mechanism of catalytic domain of human vascular endothelial growth factor receptor-2 tyrosine kinase (VEGFR2 TK), a key enzyme in angiogenesis

Vascular endothelial growth factor (VEGF) is a dimeric protein which induces formation of new blood vessels (angiogenesis) through binding to VEGF-receptor-2 tyrosine kinase (VEGFR2 TK) or KDR (kinase insert domain-containing receptor) on the surface of endothelial cells. Angiogenesis has been shown to be essential for malignancy of tumors; therefore, VEGFR2 TK is a potential therapeutic target for the treatment of cancer. Sequence homology studies indicate that VEGFR2 TK contains three domains: extracellular (ligand-binding domain), transmembrane, and intracellular (catalytic domain). In this work, the catalytic domain of VEGFR2 TK was cloned and expressed in a soluble active form using a baculovirus expression system. In the absence of ligand, the enzyme is shown to catalyze its autophosphorylation in a time-dependent and enzyme-concentration-dependent manner, consistent with a trans mechanism for this reaction. Mass spectrometry analysis revealed incorporation of 5.5 +/- 0.5 mol of phosphate/mole of enzyme (monomer). In addition, the enzyme was shown to catalyze phosphorylation of a synthetic peptide, poly(E4Y). Using poly(E4Y) as substrate, the kinetic constants of both native and phosphorylated enzyme were determined. Enzyme phosphorylation increased catalytic efficiency of the enzyme by at least an order of magnitude. Furthermore, the enzyme was shown to catalyze the reverse reaction using phospho-poly(E4Y) as substrate. Cd2+ was found to be an inhibitor of the enzyme. Kinetic studies revealed that inhibition by Cd2+ was competitive with respect to Mg2+ and noncompetitive with respect to MgATP. These results indicate that Cd2+ competes for a second metal-binding site. Therefore, the reaction catalyzed by this enzyme was treated as a terreactant system. The kinetic mechanism of VEGFR2 TK was elucidated through the use of steady-state kinetic studies. According to these studies, the enzyme binds Mg2+ and MgATP in a random fashion followed by ordered addition of the peptide substrate. The release of product is also ordered, with MgADP being released last. The order of substrate binding was confirmed by using AMP-PCP, a dead-end inhibitor.

Growth factors in the regenerating pancreas of gamma-interferon transgenic mice

We examined the distribution of several relevant growth factors in gamma-interferon transgenic mice, which undergo continual growth and differentiation in the pancreas. As a result, epidermal growth factor (EGF), TGF-alpha, and the EGF receptor were identified as potentially important in mediating some of these regenerative changes. Transient up-regulation of EGF, TGF-alpha, and the EGF receptor were observed in acini undergoing differentiation into duct-like structures. These ducts have been shown to proliferate and potentiate regeneration of the pancreatic islet mass.

Preparation and characterization of a bifunctionally spin-labeled mutant of murine epidermal growth factor for saturation-transfer electron paramagnetic resonance studies of the growth factor/receptor complex

In this report we describe the production of a [Lys3,Tyr22] murine epidermal growth factor (mEGF) mutant for spin-labeling with bis(sulfo-N-succinimidyl)-[15N,2H16]doxyl-2-spiro-4'-pimelate ([15N,2H16]BSSDP) in order to study the rotational dynamics of the EGF/EGF receptor complex by saturation-transfer electron paramagnetic resonance (ST-EPR). Previous results [Faulkner-O'Brien et al. (1991) Biochemistry 30,8976-8985] indicated that the reaction of [15N,2H16]BSSDP with wild-type mEGF did not yield a product useful for ST-EPR studies of the EGF/EGF receptor complex because the major product, in which [15N,2H16]BSSDP was attached only at the amino terminus of mEGF, lacked rigid motional coupling of the spin probe to the protein, and the more tightly coupled bidentate product was unstable. Using oligonucleotide-mediated site-directed mutagenesis of a synthetic gene for mEGF, we replaced Tyr3 with Lys and His22 with Tyr in wild-type mEGF to produce a mutant mEGF suitable for [15N,2H16]BSSDP labeling. The [Lys3,Tyr22] mEGF was expressed in Escherichia coli HB101 transformed with a pIN-III-ompA3-[Lys3,Tyr22] mEGF plasmid and was purified from the bacterial periplasm using a simple two step purification method. The [15N,2H16]BSSDP reacted with [Lys3,Tyr22]mEGF in high yield, and EPR analysis of the major product revealed tight motional coupling between the spin probe and the protein. Biological activity, as assessed by stimulation of EGF receptor autophosphorylation and dimerization, was not affected by either the mutations or the addition of the spin label. The [15N,2H16]BSSDP-modified [Lys3,Tyr22] mEGF was shown to be equipotent with mEGF in EGF receptor competition binding assays using A431 cells; in EPR studies, mEGF also was shown to specifically block [15N,2H16]BSSDP-modified [Lys3,Tyr22]mEGF binding to the EGF receptor in A431 membrane vesicles. Using the [15N,2H16]BSSDP-modified [Lys3,Tyr22]mEGF, we now report the first measurement of the rotational dynamics of the EGF/EGF receptor complex in A431 membrane vesicles by ST-EPR.

Identification of biologically active epidermal growth factor precursor in human fluids and secretions

A biologically active form of the epidermal growth factor (EGF) precursor has been detected in human fluids and secretions. The secreted protein identified in human urine and milk has an apparent molecular mass of 160-170 kilodaltons and exhibits an affinity for the glycosaminoglycan heparin. More importantly, the secreted EGF precursor is capable of activating the intrinsic tyrosyl kinase activity of the EGF receptor. Our results demonstrate that the soluble form of the precursor is generated from the membrane-anchored form by a processing step that takes place at the cell surface and involves truncation of the cytoplasmic and transmembrane domains of the intact EGF precursor. The findings support the hypothesis that the secreted 160- to 170-kilodalton EGF glycoprotein that accumulates in urine and milk is proteolytically derived from the plasma membrane-spanning precursor expressed in the kidney and mammary gland.

Expression of epidermal growth factor precursor cDNA in animal cells

This chapter outlines in detail the optimal conditions for the expression of human recombinant proteins in mouse cells using a bovine papillomavirus-based mammalian expression vector. The procedures we have described were used to successfully express high levels of the human EGF precursor in our laboratory and the human insulin receptor in the laboratory of Whittaker. Using this experimental approach we were able to demonstrate that expression of a cDNA for prepro-EGF produces a glycosylated membrane protein with biological activity.

Recombinant human epidermal growth factor precursor is a glycosylated membrane protein with biological activity

NIH 3T3 cells were transfected with cDNA corresponding to human kidney prepro-epidermal growth factor (preproEGF) under control of the inducible mouse metallothionein promoter. The synthesis of recombinant human EGF precursor by these cells has provided us with a model system for analysis of the structure and activity of this precursor. In transfected cells, the precursor was present as an intrinsic 170-kilodalton membrane protein as well as a soluble protein in the extracellular medium; both forms were N glycosylated. Glycosylation of the EGF precursor was determined by (i) the direct incorporation of [3H]mannose and [3H]glucosamine, (ii) metabolic labeling in the presence or absence of glycosylation inhibitors, (iii) enzymatic cleavage of the precursor by N-glycanase or endoglycosidase II, and (iv) lectin chromatography. Recombinant human preproEGF was purified by affinity chromatography, using wheat germ lectin and antibodies to human EGF. The intact precursor was biologically active. Purified preparations of preproEGF (i) competed with 125I-labeled EGF for binding to the EGF receptor in intact fibroblast cells, (ii) activated the intrinsic tyrosine kinase activity of the EGF receptor in membrane preparations, and (iii) sustained the growth of a mouse keratinocyte cell line that is dependent on EGF for growth. These results suggest that proteolytic processing of the precursor may not be essential for its biological function.

Differentiation of HL-60 cells is associated with an increase in the 35-kDa protein lipocortin I

Lipocortin I, a 35-kDa protein, has been detected in terminally differentiated monocytes and neutrophils. This calcium-phospholipid binding protein appears to be identical to a 35-kDa protein that can serve as a substrate for the EGF-receptor/tyrosine kinase. We have used the human myelocytic cell line HL-60 to explore whether differentiation of hematopoietic cells is associated with changes in the level of lipocortin I. We find that differentiation of HL-60 cells toward the macrophage lineage by the addition of phorbol esters or vitamin D3 or toward neutrophils with dibutyryl cyclic AMP or dimethyl sulfoxide is accompanied by an increase in the cellular content of lipocortin I. In comparison, treatment of HL-60 cells with bryostatin 1, a compound that activates protein kinase C but does not differentiate HL-60 cells, did not effect the level of 35 kDa protein. We have developed a radioimmunoassay to quantitate this protein by using a polyclonal antibody to a synthetic amino terminal peptide of the 35-kDa protein. This antibody recognizes purified pig lung 35-kDa protein as well as a single 35-kDa protein in HL-60 and A-431 cells as determined by Western blotting and immune precipitation. Differentiated HL-60 cells contain 2.6-fold the amount of 35-kDa protein found in undifferentiated HL-60 cells. Our findings that the addition of phorbol esters to HL-60 cells results in an increase in the mRNA for the 35-kDa protein and in an increase in the incorporation of 35S-methionine into the protein suggest that transcriptional activation or increased stability of the mRNA is responsible for the increased rate of synthesis and accumulation of lipocortin I during differentiation of these cells. In the absence of added divalent cations, we have determined that in differentiated HL-60 cells 79% of lipocortin I protein is located in the cytosol while 21% of the total cellular protein is bound to the particulate fraction. The 35-kDa protein can be removed from the particulate fraction by incubation with chelators or treatment with phospholipase A2 or phospholipase C. Addition of the calcium ionophore A23187 to intact differentiated HL-60 cells causes the 35-kDa protein to associate with the particulate fraction of the cell, suggesting that modulation of intracellular calcium levels may play a role in changing the intracellular location of this protein.

Expression of human epidermal growth factor precursor cDNA in transfected mouse NIH 3T3 cells

Stable cell lines expressing the human epidermal growth factor (EGF) precursor have been prepared by transfection of mouse NIH 3T3 cells with a bovine papillomavirus-based vector in which the human kidney EGF precursor cDNA has been placed under the control of the inducible mouse metallothionein I promoter. Synthesis of the EGF precursor can be induced by culturing the cells in 5 mM butyric acid or 100 microM ZnCl2. The EGF precursor synthesized by these cells appears to be membrane associated; none is detectable in the cytoplasm. The size of the EGF precursor expressed by these cells is approximately 150-180 kDa, which is larger than expected from its amino acid sequence, suggesting that it is posttranslationally modified, presumably by glycosylation. The EGF precursor was also detected in the conditioned medium from these cells, indicating that some fraction of the EGF precursor synthesized by these transfected cells may be secreted. Preliminary data suggest that this soluble form of the EGF precursor may compete with 125I-labeled EGF for binding to the EGF receptor. These cell lines should be useful for studying the processing of the EGF precursor to EGF as well as determining the properties and possible functions of the EGF precursor itself.

A calcium-dependent 35-kilodalton substrate for epidermal growth factor receptor/kinase isolated from normal tissue

We have previously reported the isolation of a 35-kDa protein from A-431 cells that, in the presence of Ca2+, can serve as a substrate for the epidermal growth factor (EGF) receptor/tyrosine kinase (Fava, R.A., and Cohen, S. (1984) J. Biol. Chem. 259, 2636-2645). We now report the detection of an antigenically related 35-kDa protein in a number, but not all, of rat, pig, and human tissues. These antigenically related proteins also can serve as substrates for the EGF receptor/kinase in the presence of Ca2+. All of these proteins share the property of reversible, Ca2+-dependent binding to the particulate fraction (presumably membranes) of cell homogenates. We have isolated the 35-kDa substrate from porcine lung and have demonstrated that it is a Ca2+-binding protein. The amino-terminal sequence and the site of tyrosine phosphorylation therein have been determined. The positions of the acidic amino acid residues amino-terminal to the tyrosine phosphorylation site bear a distinct resemblance to the sequence in the homologous region of a number of other substrates for tyrosine kinases. Based on available data, the 35-kDa protein clearly differs from the protein I complex derived from intestinal mucosa and thought to be related to the proteins isolated herein (Gerke, V., and Weber, K. (1985) J. Biol. Chem. 260, 1688-1695). Finally, we report a striking sequence homology between the porcine 35-kDa described herein and human lipocortin, a phospholipase A2 inhibitor.

ATP-stimulated interaction between epidermal growth factor receptor and supercoiled DNA

The receptor for epidermal growth factor (EGF) has been identified as a transmembrane glycoprotein that has tyrosine-specific kinase activity. The kinase activity of the receptor is enhanced in the presence of EGF (or related peptides), and the phosphorylation of a number of substrates, as well as autophosphorylation of the receptor, has been reported. Analogous findings have been described for the insulin receptor and the receptor for platelet-derived growth factor (PDGF). Thus, a number of hormone receptors and several viral transforming proteins appear to share the highly unusual property of tyrosine-specific kinase activity. Nevertheless, the specific relationship between tyrosine kinase activity and the control of cell growth and replication is unknown. It is known that after the initial binding of EGF to the plasma membrane, the hormone together with its receptor is rapidly internalized in endocytic vesicles and the hormone is eventually degraded in lysosomes. It is possible that the function of EGF is simply to stimulate internalization of its receptor, and that as a result of its altered location the receptor is able to phosphorylate a cytoplasmic component or even interact directly with a nuclear component. We now report that the purified receptor for EGF is able to interact with and nick supercoiled double-stranded DNA in an ATP-stimulated manner.

Characterization of translational-control ribonucleic acid isolated from embryonic chick muscle

Myosin heavy chain (MHC) mRNP particles have been purified from 13-day chick embryonic skeletal muscle by a combination of sucrose density gradient centrifugation and metrizamide buoyant density centrifugation. Associated with the mRNPs are at least three distinct low molecular weight RNA molecules including translational-control RNA (tcRNA). This particular RNA contains 102 nucleotides and is uridine and guanine rich, and its nucleotide sequence has been determined. tcRNA102 is capable of inhibiting the translation of the mRNAs with which it is associated upon preincubation in stoichiometric amounts. Under these conditions, endogenous reticulocyte mRNA is not inhibited. Under appropriate salt and temperature conditions, tcRNA102 is capable of reassociating with myosin heavy chain (MHC) mRNA, thus altering the sedimentation characteristics of the mRNA. This suggests that the mRNA-tcRNA102 interactions alter the secondary structure of the mRNA. In addition, tcRNA102 does not associate with ribosomal RNA or globin mRNA, suggesting that some degree of specificity is involved with the RNA-RNA interactions.

Uptake and utilization of mRNA by myogenic cells in culture

Primary chick myoblast cultures demonstrate the ability to take up exogenously supplied polyadenylated RNA and express the encoded information in a specific manner. This expression is shown to exhibit tissue specificity. Analysis of creatine kinase activity monitored at various times of incubation in the presence of either polyadenylated or nonpolyadenylated RNA indicates that only the poly(A)+ mRNA is capable of being actively translated. Radioactively labled poly(A)+ mRNA is taken up by the cell cultures in a time-dependent manner and subsequently shown to be associated with polysomes. This association with polysomes does not occur in the presence of puromycin and is unaffected by actinomycin D. Thus, nonspecific interaction with polysomes and induction of new RNA synthesis are ruled out and the association of the exogenously supplied poly(A)+ mRNA with polysomes is indicative of its translation in the recipient cells. When heterologous mRNA (globin) is supplied to the myoblasts, it is also taken up and properly translated. In addition, exogenously supplied myosin heavy chain mRNA is found associated with polysomes consisting of 4-10 ribosomes in myoblast cell cultures while in myotubes it is associated with very large polysomes, thus reflecting the different translational efficiencies that this message exhibits at two very different stages of myogenesis. The results indicate that muscle cell cultures can serve as an in vitro system to study translational controls and their roles in development.