Overlapping genes coded in the 3′-to-5′-direction in mitochondrial genes and 3′-to-5′ polymerization of non-complementary RNA by an ‘invertase’

Suppressor tRNAs induce expression of additional (off-frame) genes coded by stopless genetic codes without lengthening genomes, decreasing DNA replication costs. RNA 3′-to-5′ polymerization by tRNAHis guanylyltransferase suggests further cryptic code: hypothetical ‘invertases’ polymerizing in the 3′-to-5′ direction, advancing in the 5′-to-3′ direction would produce non-complementary RNA templated by regular genes, with different coding properties. Assuming ‘invertase’ activity, BLAST analyses detect GenBank-stored RNA ESTs and proteins (some potentially coding for the hypothesized invertase) for human mitochondrial genes. These peptides’ predicted secondary structures resemble their GenBank homologues’. 3′-to-5′ EST lengths increase with their self-hybridization potential: Single-stranded RNA degradation perhaps limits 3′-to-5′ elongation. Independent methods confirm predicted 3′-to-5′ overlapping genes: (a) Presumed 3′-to-5′ overlapping genes avoid codons belonging to circular codes; (b) Spontaneous replicational deamination (mutation) gradients occur at 3rd codon positions, unless these are involved in overlap coding, because mutations are counter selected in overlapping genes. Tests a and b converge on predicted 3′-to-5′ gene expression levels. Highly expressed ones include also fewer stops, and mitochondrial genomes (in Primates and Drosophila) adapt to avoid dependence of 3′-to-5′ coding upon antitermination tRNA activity. Secondary structure, circular code, gradient and coevolution analyses yield each clear positive results independently confirming each other. These positive results (including physical evidence for 3′-to-5′ ESTs) indicate that 3′-to-5′ coding and invertase activity is an a priori improbable working hypothesis that cannot be dismissed. Note that RNAs produced by invertases potentially produce triple-stranded DNA:RNA helices by antiparallel Hoogsteen pairings at physiological pH, as previously observed for mitochondrial genomes.

► 3′-to-5′ RNA is polymerized in the 5′-to-3′ direction by a hypothetical invertase. ► Self-hybridization by the nascent 3′-to-5′ polymerized RNA promotes its elongation. ► 3′-to-5′ RNA includes numerous undetected overlapping protein coding genes. ► These code for DNA binding proteins, potentially including the hypothesized invertase.