The Genetic Code and RNA-Amino Acid Affinities

Abstract

A significant part of the genetic code likely originated via a chemical interaction, which should be experimentally verifiable. One possible verification relates bound amino acids (or perhaps their activated congeners) and ribonucleotide sequences within cognate RNA binding sites. To introduce this interaction, I first summarize how amino acids function as targets for RNA binding. Then the experimental method for selecting relevant RNA binding sites is characterized. The selection method's characteristics are related to the investigation of the RNA binding site model treated at the outset. Finally, real binding sites from selection and also from extant natural RNAs (for example, the Sulfobacillus guanidinium riboswitch) are connected to the genetic code, and by extension, to the evolutionary progression that produced the code. During this process, peptides may have been produced directly on an instructive amino acid binding RNA (a DRT; Direct RNA Template). Combination of observed stereochemical selectivity with adaptation and co-evolutionary refinement is logically required, and also potentially sufficient, to create the striking order conserved throughout the present coding table.

Keywords: DRT; anticodon; binding; codon; triplet.

Figure 1

Amino acids studied by selection. The drawing divides the amino acids into two sites of possible interaction, divided by a horizontal dashed line. Firstly, α-carbon groups, which can make favorable polar interactions with RNA in every case. Then side chains, which can make stabilizing RNA contacts when their varied polar character allows. Val = valine; Ile = isoleucine; Leu = leucine; Gln = glutamine; Phe = phenylalanine; Tyr = tyrosine; Trp = tryptophan; His = histidine; Arg = arginine. His imidazole is drawn protonated, because the major His site [3] and the His-Phe peptide site as well [4] prefer protonated His imidazole.

Figure 2

Probability of uniform distribution of codon and anticodon triplets. Here the Implicated Site Nucleotides (ISN) are compared with other initially randomized positions in individual selected amino acid binding RNAs. Fixed sequences, of course, are not considered. Under the null hypothesis that cognate triplets (listed in columns) are equally frequent inside and outside the ISN, the probabilities of equal triplet distributions for eight kinds of amino acid sites are tabulated. Probabilities come from a two-tailed G test with Williams correction [39]. Probability boxes containing dashes are triplets that did not exist in the experimental sample. Among probabilities, shaded boxes with italicized numbers are significant. To evaluate significance in a conservative way, I compute P_sig = 1 − (1 − P_err)^1/n where P_sig is the maximum acceptable probability and P_err is the target error for each of the n trials in the Figure. To limit the probability of error to P_err = 0.01 in 44 individual trials, the maximum probability regarded as significant is P_sig = 2.3 × 10⁻⁴. Among triplets, italic triplets on white backgrounds are those concentrated by sequential squeezed selections for the cognate amino acid. Binding and sequence data can be found in: Ile [8,20,25], leucine (Leu) (I. Majerfeld, M. Illangasekare, M. Yarus, unpublished; see [1], Gln (C. Scerch and G. Tocchini-Valentini, pers. comm; see [1]), Phe [40], Tyr [41], Trp [28,29], His [3,35], Arg [9,22,26,42].

Figure 3

An example: the most prevalent Arg-binding RNA. Arg-606 [9], derived from a 25-nucleotide randomized region, is shown. Lower-case letters are fixed sequences, capital letters represent originally randomized nucleotide positions. The nucleotide sequence is threaded through the probable secondary structure for all related isolates, deduced by BayesFold [43]. Gray circles mark Implicated Site Nucleotides, and the three open gray circles are a very highly conserved arginine anticodon (cognate to codon AGG). Arg-606 had K_D = 0.5 mM, and D/L ≈ 35, consistent with the idea that the smallest sidechain-specific sites are predominantly single-ended. Comparable simplest His, Ile, and Trp sites from separate sequentially squeezed selections have been reviewed [27].

Figure 4

Schematic structure of a Sulfobacillus guanidinium riboswitch. Gray numbered circles are nucleotides of the crystallographic structure for guanidinium ion bound to the sensor of the guanidinium-I riboswitch. Dotted lines are hydrogen bonds; the gray and black curved arrows indicate that G72 covers the top, and G88 forms the bottom of the guanidinium binding site, respectively. Arg triplet nucleotides are colored; the ones centered at G45 and G72 are almost completely conserved; at G88, ≈75% conserved. G90 (black) is a non-triplet site nucleotide. Drawn from [51] and Protein Data Base (PDB) structure 5T83.

Figure 5

RNA 8, which has affinity for His and Phe in His-Phe. Lower case letters are fixed sequences, capital letters are initially randomized positions [4]. The RNA is threaded through the most probable secondary structure computed for its independent isolates by BayesFold [43]. Gray circles mark Implicated Site Nucleotides; those with white centers are potential coding triplets labeled “Phe ac” (ac = anticodon) and “His ac”. Green nucleotides are non-site, but initially randomized nucleotides.