Research Article: The Fate of miRNA* Strand through Evolutionary Analysis: Implication for Degradation As Merely Carrier Strand or Potential Regulatory Molecule?

Date Published: June 30, 2010

Publisher: Public Library of Science

Author(s): Li Guo, Zuhong Lu, Jason E. Stajich.

Abstract: During typical microRNA (miRNA) biogenesis, one strand of a ∼22 nt RNA duplex is preferentially selected for entry into a silencing complex, whereas the other strand, known as the passenger strand or miRNA* strand, is degraded. Recently, some miRNA* sequences were reported as guide miRNAs with abundant expression. Here, we intended to discover evolutionary implication of the fate of miRNA* strand by analyzing miRNA/miRNA* sequences across vertebrates.

Mature miRNAs based on gene families were well conserved especially for their seed sequences across vertebrates, while their passenger strands always showed various divergence patterns. The divergence mainly resulted from divergence of different animal species, homologous miRNA genes and multicopy miRNA hairpin precursors. Some miRNA* sequences were phylogenetically conserved in seed and anchor sequences similar to mature miRNAs, while others revealed high levels of nucleotide divergence despite some of their partners were highly conserved. Most of those miRNA precursors that could generate abundant miRNAs from both strands always were well conserved in sequences of miR-#-5p and miR-#-3p, especially for their seed sequences.

The final fate of miRNA* strand, either degraded as merely carrier strand or expressed abundantly as potential functional guide miRNA, may be destined across evolution. Well-conserved miRNA* strands, particularly conservation in seed sequences, maybe afford potential opportunities for contributing to regulation network. The study will broaden our understanding of potential functional miRNA* species.

Partial Text: MicroRNAs (miRNAs) are an abundant class of small non-protein-coding RNAs that have emerged as key post-transcriptional regulators of gene expression in animals and plants [1], [2]. Metazoan miRNA genes are transcribed by either RNA polymerase II or RNA polymerase III into primary miRNA transcripts (pri-miRNAs) as single genes or in clusters [1], [3], [4], [5]. The pri-miRNAs contain stem-loop structures (hairpins) that harbor the miRNAs in the 5′ or 3′ half of the stem. These primary miRNA gene transcripts are typically, but not always, recognized and cut by the endonuclease Drosha in the cell nucleus to produce miRNA hairpin precursors that are then exported to the cytosol, where the hairpin structures are cut by the endonuclease Dicer at relatively fixed positions and released as short double-stranded RNA duplexes [6], [7], [8], [9], [10], [11], [12]. Although both strands of duplexes are necessarily produced in equal amounts by transcription, their accumulation is asymmetric at steady state [13]. Based on the thermodynamic stability of each end of this duplex, one of the strands is thought to be a biologically active miRNA, and the other is considered as an inactive strand and a carrier strand called miRNA* (miRNA star) or passenger strand [14]. Generally, the miRNA* strand is typically degraded, whereas the mature miRNA strand is taken up into the microribonucleoprotein complex (miRNP) [6] (Figure 1A and Figure 1B). The mature miRNA strand is used as a guide to direct negatively post-transcriptional regulation by the binding of 5′-seed (nucleotides 2–8) and anchor (nucleotides 13–16) with target sequences in the 3′ untranslated region (UTR) of cognate mRNAs [1], [15]. Once bound to Ago proteins, miRNAs are more stable than average mRNAs and the half-life of most miRNAs is greater than 14 hours [16]. They may be produced by 5′ (left arms) or 3′ arms (right arms) of the miRNA precursors, and the nonrandom nature of miRNA strand selection might reflect an active process that minimizes the population of silencing complexes with illegitimate miRNA* species [13] (Figure 1). The mechanism of strand selection maybe correlates with the relative free energies of the duplex ends [11], [13], [17].

Mature miRNAs (miR-#-5p or miR-#-3p) were evolutionarily conserved across the animal kingdom [26], [27], [28], while their passenger strands, either as typically degraded miRNA* or abundant mature miRNAs, always showed conservation across vertebrates with higher nucleotide divergences than their partners (Figure 2). Different miRNA* sequences showed various divergence patterns. Data analysis revealed that some mature miRNAs and their passenger strands were well-conserved, especially in their seed sequences (Figure 2). For example, miR-124 (Figure 2A), a phylogenetic conserved miRNA from Caenorhabditis to Homo, is one of the most abundantly expressed miRNAs in the nervous system and contributes to the development of nervous system [36], [37], [38]. However, some miRNA* sequences showed higher level of nucleotide divergence, even though in their positions 2–8, such as miR-10* and miR-100* (Figure 2B). Even if multicopy hairpin precursors could yield the same mature miRNA sequences, another product, termed as miRNA* species, always diverged, such as hsa-let-7a-1* and hsa-let-7a-2*, hsa-let-7f-1* and hsa-let-7f-2* (Figure 3A). Despite of greater divergence than miRNAs, we also found miRNA* diverged much more slowly than terminal loops, which maybe strongly aid the identification of functional animal miRNA hairpins as “saddle” structures [39], [40]. Interestingly, similar divergence trends of human let-7/let-7* could be detected by sequence analysis across vertebrates (Figure 3), which might reveal historical miRNA gene divergence and similar evolutionary trend across different animals. High divergence levels could be detected among these homologous miRNA genes (Figure 3 and Figure 4). Although the divergence mainly resulted from the loop regions, the divergence of miRNA* strands also contributed partly to the high divergence level (Figure 3A and Figure 4). On the other hand, we selected several typical vertebrate animals to analyze miRNA/miRNA* sequences because different miRNAs had different distribution spectrums across the animal kingdom. Similarly, some miRNA* strands were highly conserved, but others were less conserved despite their mature miRNAs were well conserved (Figure 5). Those miRNAs that reported both miR-#-5p and miR-#-3p could be mature functional miRNAs always were well conserved especially for seed and anchor sequences, such as miR-17, miR-140 and miR-455 (Figure 5). Nevertheless, some miRNA* strands were diverged even in their seed sequences, such as miR-31*, miR-100* and miR-125b* (Figure 5). Therefore, across miRNA gene evolution, functional mature miRNAs still were well conserved especially for their seed sequences, while miRNA* sequences showed various evolutionary patterns. Some miRNA* maybe showed high divergence levels between different precursors even between different multicopy precursors, but others ensured well-conserved seed sequences, especially for those miRNA genes that generated abundant miRNAs from two arms of hairpins (Figure 2, Figure 5 and Figure S1). Evolutionary conservation of passenger strand might result from two plausible reasons. Firstly, evolutionary process would be influenced because it maybe contributed to stable stem-loop structure of miRNA hairpin precursor. Secondly, the well conservation of passenger strand might afford an opportunity to be mature miRNA to bind target mRNA similar to its partner. Therefore, the evolutionary patterns of miRNA* might be a pivotal implication (discussed below).

All the miRNA and miRNA* sequences, and their miRNA precursor sequences from different animal species were obtained in miRBase database (version 14.0, We denoted the miRNA precursors by mir-#, the mature miRNAs by miR-#, and miRNA* (miRNA star) by miR-#* in accordance with the convention in miRBase database. If the miRNA* strands were reported as abundant mature miRNA, miR-#-5p or miR-#-3p was denoted. In the study, miR-#-5p and miR-#-3p were identified according to human miRNAs in miRBase database. These sequences were aligned with Clustal X 2.0 [42] by using the multiple sequence alignment. Phylogenetic network of miRNA genes was reconstructed using the neighbor-net method [43] based on Jukes-Cantor model as implemented in SplitsTree 4.10 [44]. For human let-7 family, we attempted to reconstruct the evolutionary history from the gene tree and discover potential evolutionary implications of let-7 and let-7*. All the gaps/missing data were deleted in phylogenetic network.