No abstract

The platelet-derived growth factor (PDGF) binds via a defined amino acid sequence to heparin (Fager et al., 1992, In Vitro Cell. Dev. Biol., 28A:176-180) and the protein moiety of low density lipoproteins (LDL; apo B-100) via a similar sequence to chondroitin sulfate (Camejo et al., 1988, Arteriosclerosis Thromb., 8:368-377). In this study, synthetic oligopeptides were used to explore the capacity of smooth muscle cell-derived glycosaminoglycans to bind to the critical sequences of PDGF and apo B-100. In vitro, proliferating human arterial smooth muscle cells synthesized twice as much proteoglycans as did quiescent cells. The dominating glycosaminoglycan side chains were chondroitin and heparan sulfates in secreted and cell-associated proteoglycans, respectively. The chondroitin sulfate-rich proteoglycans had a higher molecular size and were to a larger extent secreted into the culture medium than the heparan and dermatan sulfate-rich proteoglycans. Heparan, dermatan, and chondroitin sulfates bound to the PDGF-derived oligopeptide with affinities similar to those of heparin. However, while heparan and dermatan sulfates both inhibited DNA synthesis in human arterial smooth muscle cells, chondroitin sulfate had no such inhibitory effect. Like the PDGF-derived oligopeptide, the apo B-100-derived oligopeptide bound to these glycosaminoglycans. At the same time, both oligopeptides displaced bound LDL from chondroitin sulfate in vitro and released the block on DNA synthesis in smooth muscle cells that heparin induced in culture. Thus, chondroitin, dermatan, and heparan sulfates produced by arterial smooth muscle cells may bind LDL and PDGF competitively in atherogenesis but only heparan and dermatan sulfates inhibit cellular DNA synthesis. LDL and PDGF deposition may occur by binding to similar binding sites on glycosaminoglycans derived from smooth muscle cells within atherosclerotic lesions.

We have investigated the influence of structures in the tRNA anticodon loop and stem on the ability of the anticodon to discriminate among codons. We had previously shown that anticodon UCC, when placed in the structural context of tRNAG?y from Escherichia coli, discriminated efficiently between the glycine codons, as required by the wobble rules. Thus, this anticodon read GGA and GGG but did not read GGU and GGC, whereas in mycoplasma tRNAGIY, the same anticodon did not discriminate among the glycine codons. We have now determined the reading properties of three constructions based on tRNAlGly containing the anticodon UCC in different structural contexts. In one of these constructs, tRNAlIY-ASL, the anticodon loop and stem are the same as in mycoplasma tRNAGIY. The second construct, tRNAG,'-AS, has an anticodon stem identical with the mycoplasma tRNAGIy, whereas in the last construct, tRNAGly-C32 the only difference from tRNA?GY(UCC) is that the uridine in position 32 of the anticodon loop has been replaced by cytidine. These constructs were tested for ability to read glycine codons in an in vitro protein-synthesizing system that allowed us to monitor separately the reading of each codon. We found that the anticodon UCC, when present in tRNAG'Y-AS, discriminated among the glycine codons, whereas in the constructs tRNAGI1-ASL and tRNA ¶'Y-C32, the same anticodon had lost its ability to discriminate-i.e., it behaved as in mycoplasma tRNAGIY. These results strongly suggest that nt 32 of the anticodon loop of tRNAO'Y(UCC) decisively influences the reading properties of the anticodon UCC.Analysis of tRNA genes and their gene products in mitochondria (1-3) has revealed that in these organelles each family box (group of four codons that have the same nucleotides in the first two positions and that all denote the same amino acid) is read by only one tRNA, which cannot discriminate among the nucleotides in the third-codon position, as required by the wobble rules (4). A similar situation exists in chloroplasts (5, 6), although in this case only about half of the codon families are read without discrimination. Our own analysis of Mycoplasma mycoides indicated that in this microorganism the family boxes, except for the threonine box, are read by only one tRNA each (7-9). The same results have been obtained in Mycoplasma capricolum by Andachi et al. (10). tRNAs that read without discrimination among nucleotides in the third-codon position, in most cases, have an unsubstituted uridine as their wobble nucleotide.These results are reminiscent of a phenomenon we had previously observed in a protein-synthesizing in vitro system programmed with the phage messenger MS2 RNA. In this system we found that family boxes could be read by a single tRNA with an undiscriminating method we referred to as "reading by two-out-of-three" (11). The capacity for this unconventional type of reading varied among different isoaccepting tRNAs. The tRNAGly(UCC) from M. mycoides (12) was almost as efficient in the unconventional reading of cod...

No abstract