Transcriptome-wide analysis of microRNA expression in the malaria mosquito Anopheles gambiae

Abstract

Background: microRNAs (miRNAs) are a highly abundant class of small noncoding regulatory RNAs that post-transcriptionally regulate gene expression in multicellular organisms. miRNAs are involved in a wide range of biological and physiological processes, including the regulation of host immune responses to microbial infections. Small-scale studies of miRNA expression in the malaria mosquito Anopheles gambiae have been reported, however no comprehensive analysis of miRNAs has been performed so far.

Results: Using small RNA sequencing, we characterized de novo A. gambiae miRNA repertoire expressed in adult sugar- and blood-fed females. We provided transcriptional evidences for 123 miRNAs, including 58 newly identified miRNAs. Out of the newly described miRNAs, 19 miRNAs are homologs to known miRNAs in other insect species and 17 miRNAs share sequence similarity restricted to the seed sequence. The remaining 21 novel miRNAs displayed no obvious sequence homology with known miRNAs. Detailed bioinformatics analysis of the mature miRNAs revealed a sequence variation occurring at their 5'-end and leading to functional seed shifting in more than 5% of miRNAs. We also detected significant sequence heterogeneity at the 3'-ends of the mature miRNAs, mostly due to imprecise processing and post-transcriptional modifications. Comparative analysis of arm-switching events revealed the existence of species-specific production of dominant mature miRNAs induced by blood feeding in mosquitoes. We also identified new conserved and fragmented miRNA clusters and A. gambiae-specific miRNA gene duplication. Using miRNA expression profiling, we identified the differentially expressed miRNAs at an early time point after regular blood feeding and after infection with the rodent malaria parasite Plasmodium berghei. Significant changes were detected in the expression levels of 4 miRNAs in blood-fed mosquitoes, whereas 6 miRNAs were significantly upregulated after P. berghei infection.

Conclusions: In the current study, we performed the first systematic analysis of miRNAs in A. gambiae. We provided new insights on mature miRNA sequence diversity and functional shifts in the mosquito miRNA evolution. We identified a set of the differentially expressed miRNAs that respond to normal and infectious blood meals. The extended set of Anopheles miRNAs and their isoforms provides a basis for further experimental studies of miRNA expression patterns and biological functions in A. gambiae.

Figure 1

Sequencing summary of small RNAs in A. gambiae . (A) Read frequency for all sequences assigned to A. gambiae and M. musculus genomes in sugar and blood-fed mosquitoes. Size distribution for the total number of sequence reads (B) and for unique sequence reads (C). The abundance of the reads between 17 and 30 nt from sugar and blood-fed mosquitoes are as indicated. (D) Read frequency for all sequences of small RNA reads. The identity and abundance of small RNA reads in sugar and blood-fed mosquitoes are as indicated.

Figure 2

Effect of Ago1 and Ago2 silencing on small RNA levels. (A) Relative expression levels of Ago1-2 and miR-989 in small RNA libraries after RNA silencing. dsRNA injection was used to deplete Ago1 and Ago2 in adult females. Relative quantity of mature miR-989 was measured by TaqMan assay and compared with miR-989 sequence read number in Ago1- and Ago2-silenced mosquitoes. Fold-changes in small RNA levels in Ago1- and Ago2-silenced mosquitoes (B, C) for miRNAs (miRBase and miRDeep2), (D, E) tRNA/rRNA/snoRNAs (Rfam); the median fold-change is shown for each plot. (F) Fold-changes in miRNA expression levels (Rfam, miRDeep2 and miRBase) normalized by snRNA U2 in Ago-silenced mosquitoes as indicated.

Figure 3

miRNA sequence variation. (A-B) 5’- and 3’-end sequence heterogeneity of the 5p- and 3p-derived reads as indicated. (C) Distribution of the predominant sequence reads grouped by their 5’-ends compared with the miRNAs reported in miRBase v19. (D) Frequency of sequence reads grouped by their 3’-end compared to the predominant miRNA read. (E) Sequence compilations of five highly abundant mosquito miRNAs showing their mature sequence including the corresponding 3’-end sequence variations. Below, the mature miRNA sequence reported in miRBase is shown in green, adjacent genomic sequence is in black. (F-G) 3‘-end extensions of the 5p- and 3p-sequences in sugar and blood-fed mosquitoes as indicated. (H) The percentage of A- and U-tailed sequence reads of the extremely abundant miRNAs.

Figure 4

Arm usage analyses in Anopheles miRNAs. (A) Proportion of sequence reads associated with the 5’ arm of miRNAs with respect to the total number of reads in sugar and blood-fed A. gambiae mosquitoes. (B) Venn diagram showing the number of putative A. gambiae miRNA homologs in D. melanogaster and Ae. aegypti. (C-D) Comparison of relative arm usage between D. melanogaster and A. gambiae (C) and between two mosquito species, A. gambiae and Ae. aegypti (D). Shown are miRs exhibiting an arm usage bias; the dashed line indicates 10-fold differences in the relative arm usage.

Figure 5

miRNA cluster analysis in A. gambiae . (A) Number of miRNAs in clusters in respect to genomic distance in A. gambiae. (B) Number of miRNAs conserved in clusters between A. gambiae and Ae. aegypti. (C) Conservation of mir-2/mir-13/mir-71 and mir-2944/mir-309/mir-286 in A. gambiae and Ae. aegypti. (D) Novel A. gambiae-specific clustering miRNA loci that generate miR (grey) and miR* (green) associated sequence reads in mosquito libraries; the predicted pre-miRNA stem-loop structures are shown below.

Figure 6

Differential expression of miRNAs in Anopheles females 3 h after regular and infectious blood feeding. (A) Fold-changes in miRNA expression are shown as a ratio of blood-fed to sugar-fed mosquitoes (p > 0.05). (B) Multiple RNA sequence alignment for mouse and human homologs of Anopheles miR-92a (miRBase v19). Sequence compilations of miR-92a mature sequence with the 3’-end variation in the mosquito libraries as indicated. (C) Sequence read abundance corresponding to two abundant classes of aga-miR-92a and mmu-miR-92a in sugar and blood-fed mosquitoes are as indicated. (D) Differentially expressed miRNAs in P. berghei infected mosquitoes; the fold-changes in miRNA expression levels are shown as a ratio of infected to non-infected blood-fed mosquito samples (p > 0.05).