The malaria transmitting mosquito identified morphologically as A. longirostris was found to comprise nine cryptic species in PNG that could be resolved by following the evolution of two molecular markers - the rDNA ITS2 and the mtDNA COI. The ITS2 is a transcribed spacer comprising part of a nuclear rDNA multicopy gene family and evolves through non-Mendelian inheritance processes [25, 26]. By following the evolution of the ITS2 it was possible to show evidence for reproductive isolation at the rDNA locus and also develop relatively simple species-diagnostic molecular tools to distinguish these putative cryptic species. The maternally inherited COI, widely used as a DNA barcode , was able to show the existence of eight divergent lineages with two genotypes (G and E) potentially obscured by the substantial intraspecific variation of genotype G and small interspecific divergence of genotype E.
The ITS2 revealed eight overt RFLP profiles with genotype discrimination apparent through single nucleotide polymorphisms and repeat insertion/deletion indels that are common to these regions. Seven of the eight RFLP genotypes generated a single heteroduplex profile with the RFLP of genotype C revealing two distinct heteroduplex profiles and thus suggesting that it contains two distinct ITS2 genotypes. Phylogenetic analysis of the cloned ITS2 and COI sequences from individuals representing each genotype revealed nine separate evolutionary units confirming all genotypes including both C1 and C2; these latter two appear as distinct sister clades. Agreement with the mtDNA COI on these species' groupings provided additional support for the ITS2 data, but this comparative approach may not be as informative for recently diverged species that do not show complete COI lineage sorting.
Despite the incomplete homogenization of the ITS2 in these mosquitoes, the spacer appears to evolve through concerted evolutionary rDNA turnover suggesting that interbreeding in reproductively isolated populations is driving different rDNA patterns that have in turn facilitated the discrimination of multiple species. It seems unlikely that the copy variants observed within these genotypes are the products of rDNA pseudogene amplification as we found little evidence of an elevated mutation rate in the conserved 5.8 S 3' gene region that makes up the first 100 nt of the ITS2 alignment. Additionally, no optimal or suboptimal ITS2 secondary structure folds (including RNA pseudoknots) could identify compensatory base changes when closely related ITS2 genotypes were examined that would have concurred with the theory that compensatory base changes in the ITS2 secondary structure correlate with sexual incompatibility [3, 7].
The mitochondrial DNA is regarded as fast evolving, but its mode of maternal inheritance can only reveal the presence of divergent lineages and cannot confirm the existence of reproductive isolation between lineages. Discrimination of species based on a percentage of divergence thresholds has been suggested , but this has been questioned for studies of Diptera . The advantage of using the rDNA as a genetic marker is that it is possible to identify reproductive isolation at the rDNA locus where different rDNA genotypes exist in sympatry. Under this assumption, most A. longirostris ITS2 genotypes were collected in sympatry with other genotypes (except H and F) while still maintaining genotype-specific RFLP and heteroduplex profiles. Additionally, most genotypes show robust phylogenetic clustering with no shared sequences that could be considered either ancestral or a result of recent hybridization. The mtDNA COI marker supports all nine genotypes with evidence of extended time in isolation in most cases. We suggest that A. longirostris is a complex of at least nine cryptic species in PNG that can now be distinguished by a relatively simple PCR-based procedures utilizing the ITS2.
The utility of the ITS2 as a marker that can show early genetic discontinuity between populations may arise from a combination of interlinked factors that include the ITS2 sequence length, the position of the rDNA array on the chromosome with an associated ability to facilitate recombination. The size of the ITS2 appears to have some bearing on the ability of the sequence to acquire non-deleterious mutations with longer sequences generally better able to accommodate mutation accumulation than shorter sequences. If indeed a core structure is necessary for maturation of the rRNA, mutations in a minimum length structure will affect the core structure integrity moreso than longer ITS2 sequences that may accommodate a higher level of mutation. For example, the extension and contractions of helices through the gain and loss of nucleotides and indels may not compromise the structure-function relationship. Using examples from Diptera, short ITS2 sequences such as those found in Nearctic Blackfly species (Diptera: Simuliidae; ITS2 length 286-260 nt) show a restricted number of informative sites when examined across five genera . The ITS2 sequences from members of the Anopheles gambiae complex (ITS2 ~ 426 nt) reveal a low level of single nucleotide polymorphisms between cryptic species [16, 29]. This study and similar studies on Anopheles mosquitoes from PNG suggest some correlation between rDNA transcribed spacer size - whether it be ITS2 or ITS1 - and the level of DNA sequence variation and homogenization, with longer spacers showing a more complex pattern of incomplete concerted evolution and DNA sequence variation that can be informative at the inter- and intraspecific level [11, 30–32].
The DNA turnover machinery responsible for this observed pattern of concerted evolution is thought to be driven primarily through unequal crossover and gene conversion [13, 33]. If recombination is reduced, as it is near the centromere , then homogenization may appear incomplete and copy variants can be carried along within an interbreeding population. The rDNA in Anopheles mosquitoes is located on the sex chromosomes adjacent to the centromeres , and so rDNA evolution would be driven primarily by intrachromosomal events, most likely gene conversion, within the rDNA array. The presence of fixed ITS2 variants within the A. longirostris rDNA genotypes suggests the rDNA array is moving as a single evolutionary unit through interbreeding populations. Comparable studies have come from work on Drosophila melanogaster, where the rDNA array exists within paracentromeric heterochromatin on sex chromosomes resulting in reduced recombination that sees rDNA turnover also occurring through intrachromosomal exchanges with the rDNA array evolving as a single evolutionary unit or locus . Intrachromosomal events driving rDNA evolution have also been suggested by others [37, 38], and the DNA turnover machinery suggested is gene conversion. Indeed biased gene conversion can be dynamic as seen in hybrid arthropod scallops where experimental hybridisation between closely related scallop species carrying different rDNA ITS variants can induce rapid early development biased gene conversion (to the maternal type) that is almost complete 14 days after fertilization .
Anopheles longirostris revealed nine independently evolving rDNA genotypes with each genotype showing fixed ITS2 sequence variants and suggesting that the rDNA array is moving as a single evolutionary unit or locus. This observation is evident in other Anopheles mosquitoes we have studied and has previously provided insights into genetic subdivisions within and between putative species [30–32]. These regions may be under divergent natural selection as evolutionary studies on A. gambiae from Africa see the X-linked centromere-proximal rDNA in a region under reduced gene flow, which has led to the suggestion that suppressed recombination in this region may be involved in speciation [40, 41]. The rDNA in A. gambiae and other Anopheles mosquitoes probably reside in regions that are now regarded as genomic islands of speciation (areas of the genome which show high levels of differentiation early in the divergence process) which often include genes responsible for behavioral or ecological isolation [42, 43]. Linkage disequilibrium between the rDNA and potential isolating factors or genes under selection may have contributed to the ITS2 evolution we see in these cryptic species from PNG.