The substrate specificity of Saccharomyces cerevisiae myristoylCoA:Protein N-myristoyltransferase: Analysis of myristic acid analogs containing oxygen, sulfur, double bonds, triple bonds, and/or an aromatic residue

We have explored the acyl-CoA substrate specificity of Saccharomyces cerevisiae myristoyl-CoA:protein N-myristoyltransferase (NMT) by synthesizing 81 fatty acid analogs and surveying their activity in a coupled in vitro assay containing Pseudomonas acyl-CoA synthetase and Escherichia coli-derived yeast NMT. Single oxygen or sulfur substitution for C-3 through C-13 is well tolerated by both enzymes. Detailed kinetic analyses suggest that the acyl-CoA and peptide-binding sites of NMT are relatively insensitive to placement of single group 6B heteroatoms. By contrast, di-oxygen-substituted analogs were very poor substrates, producing dramatic reductions in the affinity of NMTs peptide-binding site for a synthetic octapeptide substrate derived from the NH₂-terminal sequence of a known N-myristoylprotein, the gag polyprotein precursor of human immunodeficiency virus 1 (HIV-1). This observation provides an example of binding site cooperativity in NMT. Replacement of one oxygen with sulfur at either the 6, 9, or 12 position of dioxatetradecanoic acids results in a general increase in peptide catalytic efficiency (V_max/K_m). An analysis of five fatty acids from octanoic to dodecanoic having terminal phenyl groups indicated that the best substrate was 10-phenyldecanoic acid even though Corey-Pauling-Koltun molecular models indicate that it has a length equivalent to that of tridecanoic acid. Six analegs having an equivalent length of 13 carbon atoms were subsequently prepared in which the phenyl group was systematically moved one methylene group closer to carboxyl. Movement of the phenyl just one carbon closer to carboxyl. (producing 9-(p-methylphenyl) nonanoic acid) decreases peptide catalytic efficiency (V_max/K_m] severalfold compared to 10-phenyldecanoic acid. 10-(4-Tolyl)decanoic acid has the same relative positions of phenyl and carboxyl as 10-phenyldecanoic acid even though a methyl group is present on the phenyl ring. It produces peptide K_m and V_max values that are the same as 10-phenyldecanoic acid. Substilution of either oxygen or sulfur for a methylene group fails to override the effects noted when the phenyl group position is altered in the C-14 equivalent fatty acid series. Several fatty acids of differing chain lengths with cyclohexyl-, 2-furyl, and 2-thienyl groups at their ω termnius had activity profiles that paralleled those of the comparable phenyl-substituted compounds. Myristic acid analogs with triple bonds (beginning at positions 2 through 13), cis-double bonds (positions 3 through 13) and trans-double bond isomers (E5, E6, and E7) were also tested. Evaluation of the alkynyl analog series revealed that 5-tetradecynoic acid (Y5) was not a substrate for NMT even though the corresponding Y4 and Y6 compounds had activities comparable to myristate. Y5 does not appear to bind to the acyltransferase. By contrast, Z5 was a superior substrate compared to all the other 10 cis-double isomers tested. E5 was inferior to E6 and E7, both of which were comparable to myristate in acylpeptide formation. Unexpectedly, addition of Z6 to the in vitro NMT assay system yielded two acyl peptide products, one that comigrated on C₁₈ reverse-phase HPLC with Z6 and the other with purified E6. A comparable phenomenon was not observed with other Z isomers or with E6. Control experiments indicated that this apparent isomerization was not due to the Pseudomonas acyl-CoA synthetase present in the coupled assay system. Together the data suggest that the acyl chain of myristoyl-CoA is present in a bent conformation in the acyl-CoA-binding site of S. cerevisiae NMT. The site appears to possess a complex measuring device that can sense distance along the chain from carboxyl to the ω end of the fatty acid as well as steric volume at the ω terminus. The overall shape of this sensor may be conical with the cone and the terminal carbons coaxial. In this proposed model, the relative sensitivity to chain length versus steric volume would be determined by the acuteness of the "cone.".