The mosquito Aedes aegypti has a large genome size and high transposable element load but contains a low proportion of transposon-specific piRNAs

Abstract

Background: The piRNA pathway has been shown in model organisms to be involved in silencing of transposons thereby providing genome stability. In D. melanogaster the majority of piRNAs map to these sequences. The medically important mosquito species Aedes aegypti has a large genome size, a high transposon load which includes Miniature Inverted repeat Transposable Elements (MITES) and an expansion of the piRNA biogenesis genes. Studies of transgenic lines of Ae. aegypti have indicated that introduced transposons are poorly remobilized and we sought to explore the basis of this. We wished to analyze the piRNA profile of Ae. aegypti and thereby determine if it is responsible for transposon silencing in this mosquito.

Results: Estimated piRNA sequence diversity was comparable between Ae. aegypti and D. melanogaster, but surprisingly only 19% of mosquito piRNAs mapped to transposons compared to 51% for D. melanogaster. Ae. aegypti piRNA clusters made up a larger percentage of the total genome than those of D. melanogaster but did not contain significantly higher percentages of transposon derived sequences than other regions of the genome. Ae. aegypti contains a number of protein coding genes that may be sources of piRNA biogenesis with two, traffic jam and maelstrom, implicated in this process in model organisms. Several genes of viral origin were also targeted by piRNAs. Examination of six mosquito libraries that had previously been transformed with transposon derived sequence revealed that new piRNA sequences had been generated to the transformed sequences, suggesting that they may have stimulated a transposon inactivation mechanism.

Conclusions: Ae. aegypti has a large piRNA complement that maps to transposons but primarily gene sequences, including many viral-derived sequences. This, together the more uniform distribution of piRNA clusters throughout its genome, suggest that some aspects of the piRNA system differ between Ae. aegypti and D. melanogaster.

Figure 1

The size distribution and assignment of piRNAs relative to genome complexity in Ae. aegypti. A) Size distribution of Ae. aegypti small RNA abundance in a representative library (library 1). The number of small RNAs mapping to Ae. aegypti genes, transposons, both, or neither, are shown as different colors for each size class (the legend is shown on the right). B) Relative distribution of nucleotide abundance at Ae. aegypti small RNA positions for small RNAs targeted to transposons, genes and other sequences. The graph was drawn using using Weblogo [78]. C) Percentage of the sequenced Ae. aegypti genome occupied by genes, transposon, both, or neither (the legend is shown on the right).

Figure 2

Size distribution and frequency of 5' base of overlapping small RNAs (small RNAs at least 24 nt long, overlapping pairs on opposite strands) for all combined Ae. aegypti libraries (A) and the D. melanogaster library (B). The length of overlap is shown on the horizontal axes. Indicated above each axis is the number of possible overlapping pairs of small RNAs (individual small RNA sequences may be involved in multiple pairs) with specified overlap size. Indicated below each axis is the relative frequency of the 5' base identity for sequences involved in overlapping pairs. The color code for bases is indicated in the center box.

Figure 3

Presumed piRNAs from Ae. aegypti are modified at their 3' ends. Control synthetic 23-mer RNA and purified 28-32 nt RNAs from Ae. aegypti were subjected to periodate oxidation and β-elimination reactions and run on a denaturing polyacrylamide gel. Synthetic 23-mer control RNA gained mobility, as expected for an RNA not modified at its 3' end while presumed piRNAs from Ae. aegypti failed to gain mobility, consistent with these RNAs being modified at their 3' ends.

Figure 4

The abundance of piRNAs derived from five Ae. aegypti lines (libraries from replicate lines were collapsed) by size class. Each size class is divided into the number of piRNA sequences found in libraries derived from 1, 2, 3, 4, or 5 lines (the legend is shown at top right).

Figure 5

The relative location of genes, transposons, and mRNA-seq sequences inside and surrounding the top Ae. aegypti piRNA cluster. Location of the piRNA cluster is shown as a blue box near the bottom of the figure, the remaining genomic features were based on the output of the VectorBase Ae. aegypti genome browser for the region located on supercontig1.478:451736-607348.

Figure 6

Mapping locations of piRNAs to plasmid pMos3DB2Her that was used as a transformation vector for Ae. aegypti transformation. Transposon derived sequences are shown as blocks. piRNA sequences are not drawn to scale to improve legibility. piRNA sequences shown above and below the plasmid sequence represent sequences that map to the positive and negative strand respectively, as determined by the transposase ORF. No piRNAs from libraries 2, 6, 11, or 12 mapped to this sequence.

Figure 7

Location of where two Ae. aegypti piRNA sequences from library 4 map to the M. domestica Hermes transposase. Also shown is the abundance of piRNA sequences mapping to the sense and antisense strands of the Hermes transposon from a M. domestica small RNA library (abundance scales for both strands are shown on the left).

Figure 8

The abundance of D. melanogaster piRNA sequences mapping to the sense and antisense strands of the D. melanogaster tj gene (top) and Ae. aegypti piRNA sequences (all libraries combined) mapping to the Ae. aegypti homolog of the tj gene (bottom). Abundance scale is shown on the left.

Figure 9

The abundance of D. melanogaster piRNA sequences mapping to sense and antisense strands of the D. melanogaster maelstrom gene (top) and Ae. aegypti piRNA sequences (all libraries combined) mapping to the Ae. aegypti homolog of the maelstrom gene (bottom). Abundance scale is shown on the left.