Evidence for targeted gene transfer by receptor-mediated endocytosis. Stable expression following insulin-directed entry of NEO into HepG2 cells.

Evidence is presented for targeted gene delivery to HepG2 cells via the endocytotic pathway under the direction of insulin. Serum albumin was treated with the water-soluble carbodiimide N-ethyl-N'-(3-dimethylaminopropyl) carbodiimide hydrochloride and the resultant positively charged N-acylurea albumin covalently conjugated to insulin by glutaraldehyde cross-linkage. The conjugated protein is shown by nitrocellulose filter binding and gel band shift assays to bind DNA, and by competitive displacement of [125I]insulin to bind to the insulin receptor. When the expression vectors ptkNEO and pAL-8 which incorporate the neo gene were complexed to the conjugate in an in vitro system of transfection, G418 resistant clones developed at frequencies of 2.0-5.5 x 10(-5). Subsequently, a 923bp PstI fragment within the neo sequence was identified by Southern transfer in genomic DNA from transfected cell populations which had been maintained on a G418 regime for 44 days.

The preparation of poly (dT)-5'-transferrin conjugates and hybridisation studies with poly (dA)-tailed linearised pBR322 plasmid DNA.

The formation of transferrin-DNA complexes intended for ligand-directed transfection studies has been achieved through a hybridisation technique involving complementary homodeoxypolynucleotide chains attached to the participating protein and DNA species. Oligothymidylate residues (pT)n obtained by dicyclohexylcarbodiimide (CDI) polymerisation of thymidine-5'-monophosphate (5'-TMP) were activated to the 5'-imidazolides which on incubation with transferrin yielded the 5'linked phosphoramidates (pT)n-5'-transferrin. Homopolymeric chain extension of (pT)5-5'-transferrin by terminal transferase and dTTP at 30 degrees for 30 min yielded (pT) 300-5'-transferrin. Cleavage of the phosphoramide link in the polymer modified transferrin at 37 degrees was pronounced after 30 min although at 25 degrees hydrolysis was less than 5% after 4 hr. Poly(dT)-5'-transferrin readily hybridised with [3H]poly(dA)-tailed Pst 1 linearised pBR322 DNA. Resultant complexes were demonstrated by nitrocellulose filter binding and immunoprecipitation with anti-transferrin antibody. In contrast with poly(dT)-5'-transferrin, poly(dT)-5'-transferrin-poly(dA)-tailed pBR322 DNA complexes were stable at 37 degrees suggesting that annealing is followed by further stabilising interactions between the DNA and protein components.

Further studies on targeted DNA transfer to cells using a highly efficient delivery system of biotinylated transferrin and biotinylated polylysine complexed to streptavidin.

Conjugates consisting of biotinylated transferrin and biotinylated poly-L-lysine attached to streptavidin have been prepared and found to transfer luciferase plasmid DNA very efficiently to HeLa cells in the presence of chloroquine. Transfection was dependent on (i) use of biotinylated short chain polylysine containing 70 lysine residues, (ii) biotinylated transferrin containing 1-2 biotin moieties, (iii) reaction of biotinylated transferrin with streptavidin followed by isolation of the resulting conjugate on Sephadex G-200 and (iv) interaction of streptavidin-biotinylated transferrin with biotinylated polylysine giving a complex suitable for DNA transfection. It was found that if the above sequence of steps resulting in the formation of streptavidin-biotinylated transferrin/biotinylated polylysine was followed without isolation of intermediate conjugates by Sephadex G-200 chromatography, pRSVL DNA transfer was still very efficient. Transfer of luciferase DNA by the streptavidin conjugates and subsequent expression of luciferase activity was almost completely inhibited by excess free transferrin, showing that gene transfer was through the transferrin receptor pathway via receptor-mediated endocytosis. The streptavidin (bio2-transferrin) bio10-pLys70 conjugate used in the present experiments was approximately one hundred times more efficient in pRSVL DNA transfection with the HeLa cells than the previously described avidin-pLys460 (bio-transferrin) complex.

The binding of DNA to water soluble carbodiimide modified proteins.

Bovine serum albumin and human serum transferrin modified by the water soluble carbodiimides N-ethyl-N'-(3-dimethyl propylamino) carbodiimide (CDI) or N-ethyl-N'-(3-trimethyl propylammonium) carbodiimide (Me+CDI) to yield N-acylurea proteins bind pBR322 DNA reversibly showing electrostatic and non-electrostatic components in the binding energies (delta G overall). It is proposed that initially an electrostatic interaction arises from ion pair formation between the DNA phosphates and the N-acylurea entities. This is consolidated, in single stranded regions, by a second event in which it is suggested that the base guanine interacts with elements of the N-acylurea moieties through hydrogen bonding or a glyoxal-type addition.

A possible model for the methylation of deoxycytidine in DNA.

The modified base 5-methylcytidine has been found in the DNA of a number of different eukaryotic cells where it occurs principally in the dinucleotide sequence -CmpG- which is present as a palindrome in double-strand nucleic acid molecules. There is considerable evidence to indicate and suggest that 5-methylcytosine serves as a regulatory signal in eukaryotic gene expression. Replication of DNA containing -CmpG- gives rise to daughter DNA molecules containing new -CpG- dinucleotide sequences in which the cytidine residues are not methylated. Methylation of these residues is carried out by a methylase enzyme using S-adenosyl-L-methionine as a specific methyl group donor. The model discussed in the present communication tries to explain in chemical and biological terms the mechanism of the methylation reaction. The first reactions of the scheme are well known through the work of other investigators. However, we introduce a new concept into our reaction mechanism by postulating the direct involvement of S-adenosyl-L-methionine in the reaction through its covalent attachment to the cytosine ring followed by a specific ring closure and methylation involving transfer of a hydride ion. The model also gives a possible explanation of mechanism of interaction of dimethyl sulphoxide with the enzyme systems of certain eukaryotic cells, which are altered or changed in the regulation of gene expression by this chemical reagent.

Are phosphorylated tyrosine residues of certain receptors involved in inducing transitory covalent protein cross-linking?

Following ligand binding a number of cell-surface receptors become phosphorylated at tyrosine residues of their cytosolic domains. These phosphorylations are associated with initiation of a signalling programme involving a sequence of tyrosine-phosphorylated protein-protein interactions. In the recognition process between phosphorylated proteins, electrostatic interactions between negatively charged phosphorylated tyrosines, serine and threonine residues and positively charged lysines play an important role as well as hydrophobic and H-bonding reactions. We suggest in this paper that the fairly high-energy phosphate bond of certain protein phosphorylated tyrosines are possibly involved in inducing transitory protein cross-linking reactions. Through a process involving transfer of an activated phosphate of phosphorylated tyrosine to a side-chain carboxyl group of the receptor or next protein of the signalling sequence, an acyl phosphate is formed. This then acylates a hydroxyl group on a serine, threonine or tyrosine residue of the protein not carrying the carboxyl phosphate to give an ester linkage, thus cross-linking the two proteins of the signalling pathway. The covalent ester linkage is labile to hydrolysis and depending on the protein-protein molecular environment it might have a finite half-life. On hydrolysis, the transitory covalent linkage is broken with separation of the proteins. It is suggested therefore that formation of a protein-protein ester linkage introduces a type of timing device into the system. Breakdown of the original protein-phosphorylated tyrosine in this case therefore does not involve a phosphatase enzyme.

Studies on a nucleoprotein prepared from rat liver polysomes by digestion with T1 ribonuclease.

1. Treatment of rat liver polysomes in a buffer containing 2.5mm-magnesium chloride with T(1) ribonuclease at a concentration of 330units/ml. of reaction medium at 37 degrees for 2hr. leads to the production of an insoluble nucleoprotein. 2. On the bases of analysis for protein and RNA and of u.v.-absorption spectra the nucleoprotein appears to have lost approx. 60% of the structural RNA originally present in the ribosome. Degradation of (3)H-labelled polysomes (structural RNA labelled with orotic acid) with T(1) ribonuclease leads to nucleoprotein preparations retaining approx. 30% of the radioactivity originally present in the polysomes. By means of sucrose-density-gradient centrifugation it is shown that the nucleoprotein preparations are free of single 73s ribosomes and ribosomal subunits. No evidence for the presence of 28s and 18s structural RNA was obtained on examination of extracted nucleoprotein-particle RNA by means of sucrose-density-gradient centrifugation. 3. Digestion of washed polysomes carrying (14)C-labelled nascent peptide chains with T(1) ribonuclease gives a nucleoprotein particle that retains approx. 70% of the original labelled chains. Treatment of labelled nucleoprotein particles with 1mm-puromycin in the absence of transfer factors releases 20% of the labelled chains. Addition of GTP (0.48mumole) increases this release to 37%. 4. Treatment of nucleoprotein particles carrying (14)C-labelled peptide chains with either EDTA (50mm) or ammonium chloride (0.5m) brings about a small release of labelled material (approx. 15%). 5. Disruption of nucleoprotein particles carrying (14)C-labelled peptide chains with either sodium dodecyl sulphate or 2m-lithium chloride, followed by addition of transfer RNA as marker and chromatography on Sephadex G-200, show in both cases that considerable amounts of labelled peptide material move well ahead of the added transfer RNA marker. Further, if nucleoprotein particles carrying labelled peptide chains are treated with 0.3m-potassium hydroxide at 20 degrees for 24 hr., neutralized to pH7.6, and then chromatographed on Sephadex G-200, the labelled peptide material moves much closer to the added transfer RNA marker. These results suggest that a proportion of the nascent (14)C-labelled peptides on the nucleoprotein are attached to transfer RNA or large fragments of transfer RNA. 6. [(3)H]Polyuridylic acid binds to nucleoprotein particles in 1mm-magnesium chloride. The rate of binding is rapid when measured at 20 degrees .

Synthesis of thiol-containing analogues of puromycin and a study of their interaction with N-acetylphenylalanyl-transfer ribonucleic acid on ribosomes to form thioesters.

1. The thiol-containing analogue of puromycin, 6-dimethylamino-9-{1'-[3'-(2''-mercapto-3''-phenylpropionamido)-3'-deoxy-beta-d-ribofuranosyl]}purine (XVII) in which the primary amino group of the antibiotic is replaced with a thiol grouping, was synthesized chemically (compound XVII is abbreviated to thiopuromycin). 2. Thiopuromycin (XVII) was found to be active in releasing N-[(3)H]acetylphenylalanine from its tRNA carrier as the thioester, N-acetylphenylalanylthiopuromycin (XIX) in the Escherichia coli ribosomal system. The reaction product (XIX) was synthesized chemically from thiopuromycin and N-acetylphenylalanine and found to be stable to hydrolysis in the standard incubation medium at pH7.6. dl-Phenyl-lactylpuromycin (XXI), the hydroxy analogue of puromycin, was also synthesized chemically and shown to release N-acetylphenylalanine from its tRNA carrier in the E. coli ribosomal system, thus confirming the previous results of Fahnestock et al. [Biochemistry (1970) 9, 2477-2483]. 3. In marked contrast with the results obtained in the E. coli system, both thiopuromycin (XVII) and hydroxypuromycin (XXI) were found to be inactive in releasing N-acetylphenylalanine from its tRNA carrier in the rat liver ribosomal system.

Synthesis of certain puromycin analogues and their use in studying the peptidyl synthetase enzyme of e. coli and rat liver ribosomes.

The beta-alanyl and L-histidyl analogues of puromycin were synthesized chemically and tested for their ability to release [3H] N-acetylphenylalanine from its tRNA carrier in the rat liver and E. Coli ribosomal systems. Both analogues were found to be inactive in releasing [3H] N-acetylphenylalanine. Reasons for the inactivity of these compounds are discussed in relation to the structure of the puromycin molecule and the requirements for puromycin-like activity.

Synthesis of cyclohexylpuromycin and its reaction with N-acetylphenylalanyl-transfer ribonucleic acid on rat liver ribosomes.

1. Cyclohexylpuromycin, an anlogue of puromycin in which a cyclohexane ring replaces the aromatic benzene ring of the L-phenylalanyl moeity of the nucleoside., has been synthesized and examined for its ability to release N-acetylphenylalanine from tRNA attached to rat liver ribosomes. 2.dl-Cyclohexylpuromycin was active in reacting with N-[3H]acetylphenylalanyl-tRNA on rat liver ribosomes to form N-E13H]lacetylphenylalanycyclohexypuromycin. 3. The reaction product N-acetylphenylalanylcyclohexylpuromycin and the corresponding analogue N-acetylphenylalanylpuromycin were chemically synthesized for evaluation of the structure of the released N-acetylphenylalanyl-containing material. 4. The results obtained suggest that the model of Raacke (1971) for purmycin reactivity needs further examination with regard to the role played by the aromatic ring system of the Lphenylalanyl moiety of the nucleoside

Differences in the ribosomes prepared from lactating and non-lactating bovine mammary gland.

1. Ribosomes prepared from bovine lactating mammary gland are able to synthesize protein, whereas similar preparations from non-lactating glands are not. Washing the ribosome suspensions through a medium containing 0.5m-ammonium chloride enhanced their ability to incorporate phenylalanine into polyphenylalanine. 2. Ribosomes isolated from non-lactating bovine mammary gland, in contrast with those from rat liver and lactating mammary gland, contained significant amounts of extraneous nucleases. These enzymes could be removed by washing with a medium A buffer containing 0.5m-ammonium chloride. 3. Only those ribosomes from functionally active tissues were able to bind polyuridylic acid and phenylalanyl-tRNA.