Primary microRNAs (miRNAs) are the precursors of miRNAs that modulate the expression of most mRNAs in humans. They fold up into a hairpin structure that is cleaved at its base by an enzyme complex known as the Microprocessor (Drosha/DGCR8). While many of the molecular details are known, a complete understanding of what features distinguish primary miRNA from hairpin structures in other transcripts is still lacking. We develop a massively parallel functional assay termed Dro-seq (Drosha sequencing) that enables testing of hundreds of known primary miRNA substrates and thousands of single-nucleotide variants. We find an additional feature of primary miRNAs, called Shannon entropy, describing the structural ensemble important for processing. In a deep mutagenesis experiment, we observe particular apical loop U bases, likely recognized by DGCR8, are important for efficient processing. These findings build on existing knowledge about primary miRNA maturation by the Microprocessor and further explore the substrate RNA sequence-structure relationship.

初级microRNA（miRNA）是可调节人类大多数mRNA表达的miRNA的前体。它们折叠成发夹结构，该结构在其碱基处被酶复合物（微处理器（Drosha / DGCR8））切割。尽管许多分子细节是已知的，但仍缺乏对将哪些特征与其他转录物中的发夹结构区分开的主要miRNA的完整了解。我们开发了一种称为Dro-seq（Drosha测序）的大规模并行功能测定法，该方法能够测试数百种已知的主要miRNA底物和数千种单核苷酸变体。我们发现了主要的miRNA的一个附加功能，称为香农熵，它描述了对加工很重要的结构整体。在深层诱变实验中，我们观察到了可能由DGCR8识别的特定顶端U环碱基，对于有效处理很重要。这些发现建立在微处理器对初级miRNA成熟的现有知识的基础上，并进一步探索了底物RNA序列与结构的关系。