American Scientist

To the Editors:

I was excited to read Emily Mortola and Manyuan Long’s article about de novo genes arising from “junk” DNA (“Turning Junk into Us,” May–June 2021). I was surprised to learn that, in most scenarios, it follows the late open reading frame (ORF)–early transcription model. [ORFs mark the area of a gene that codes for a protein; Mortola and Long’s research found that de novo genes often start being transcribed as nonsense RNA before a complete ORF develops. See figure below for more details.]

However, the authors missed the opportunity to explain how transcription is possible without an ORF—and this being possible, why an ORF is essential (at least for “normal” genes). Furthering this idea, one might even wonder if all parts of the DNA are theoretically transcribable, but some feedback mechanism or cellular efficacy priming allows “junk” DNA to be transcribed at negligible rates and in negligible amounts, whereas ORF speeds up transcription of useful genes profusely.

Leo W. Sham
Hong Kong

Dr. Long responds:

Thank you for sharing this astute question. Although we did not spend space on the topic of widely existing intergenic transcription in this article, and chose instead to focus on the origination of de novo genes, the topic was discussed in our original research paper. We, too, were surprised when we detected that the late ORF–early transcription model is the dominant pathway for de novo origination. An ample number of previously reported intergenic noncoding RNAs (ncRNAs), especially long noncoding RNAs (lncRNAs), exist in organisms ranging from animals (including humans) to plants to fungi. These intergenic RNAs form a giant repository of almost infinite raw material from which new genes can arise.

The ORF is not required to maintain the large number of intergenic transcripts in organisms. An intergenic RNA without an ORF can be persistent in various ways. Many ncRNAs or lncRNAs have been shown to play certain important roles and thus can be maintained by purified selection. For example, the Sphinx gene in fruit flies, encoding a very long ncRNA (1.2 kilobases), was found to be involved in courtship. Genetic disruption of Sphinx was observed to lead to changes in various male courtship behaviors. Genetic knockout of 129 testis-associated lncRNAs in fruit flies through gene editing resulted in a loss of male fertility in a third of cases. In mammalian genomes, many lncRNAs were found to carry out important functions, exemplified by the lncRNA JPX gene involved in dosage compensation that silences an X chromosome in females. Meanwhile, pervasively transcribed noncoding RNAs may regulate genome expression through cis/trans mechanisms. Among plant lncRNAs, 40 percent were found to have sequence features similar to those of functional genes, whereas the remaining 60 percent were believed to be transcriptional noise. We await further studies and data in the future to shed more light on this issue.