Genetic differentiation associated with host plants and geography among six widespread species of South American Blepharoneura fruit flies (Tephritidae)

Abstract

Tropical herbivorous insects are astonishingly diverse, and many are highly host-specific. Much evidence suggests that herbivorous insect diversity is a function of host plant diversity; yet, the diversity of some lineages exceeds the diversity of plants. Although most species of herbivorous fruit flies in the Neotropical genus Blepharoneura are strongly host-specific (they deposit their eggs in a single host plant species and flower sex), some species are collected from multiple hosts or flowers and these may represent examples of lineages that are diversifying via changes in host use. Here, we investigate patterns of diversification within six geographically widespread Blepharoneura species that have been collected and reared from at least two host plant species or host plant parts. We use microsatellites to (1) test for evidence of local genetic differentiation associated with different sympatric hosts (different plant species or flower sexes) and (2) examine geographic patterns of genetic differentiation across multiple South American collection sites. In four of the six fly species, we find evidence of local genetic differences between flies collected from different hosts. All six species show evidence of geographic structure, with consistent differences between flies collected in the Guiana Shield and flies collected in Amazonia. Continent-wide analyses reveal - in all but one instance - that genetically differentiated flies collected in sympatry from different host species or different sex flowers are not one another's closest relatives, indicating that genetic differences often arise in allopatry before, or at least coincident with, the evolution of novel host use.

Anatomy of a Neotropical insect radiation

“…Nevertheless, to correct for the possibility that some divergent haplotype lineages belong to actually- or potentially-interbreeding populations, analyses were repeated using a truncated phylogeny, derived by removing lineages arising later than 1 million years ago (mega-annum, Ma), then shortening terminal branch lengths by 1 Ma. With few exceptions, analyses of intra-lineage population structure [ 24 ] and their strong geographic structure corroborate the assumption that mtCOI haplotype groups at the 1 Ma level of divergence (about 2%) are representative of real, independently evolving biological units. Both ecological and behavioral data also support the assumption that haplotype lineages at the 2% level of divergence reflect evolutionarily relevant units.…”

“…Our analyses include collections previously reported by Condon et al [ 21 – 23 ], and specimens from our more recent collections from Costa Rica, Panama, Peru, Bolivia, Brazil, Suriname, and French Guiana (Fig. 1 ; see also [ 24 ]).All specimens were collected, exported, and imported in adherence to national and international regulations, under permits issued by the relevant national agencies (see Additional file 1 ). Following methods described in Condon et al [ 21 , 23 ] we used PCR amplification and sequencing of the mitochondrial cytochrome oxidase 1 (mtCOI) gene (504 bases near the 3′ end) from adult and pupal flies using one of two sets of primers: a set of generalized primers used for adult flies (C1-J-2183/TL2-N-3014; [ 36 ], see [ 21 ]) or a set of fly-specific primers used for pre- or post-emergence puparia that avoids amplification of fragments from hymenopteran parasitoids (BpupCO1F/BpupCO1R; [ 23 ]).…”

“…Our goal in this study is to understand patterns of Blepharoneura diversification across multiple time scales and geographic dimensions. Our previous work [ 21 – 23 ] used 4% mtCOI divergence as a conservative but arbitrary cut-off to define “species.” Analyses of allozymes [ 38 ] and microsatellites [ 24 ] of sympatric populations show that members of groups (“species”) differing by 4% mtCO1 do not interbreed. Furthermore, microsatellite data [ 24 ] clearly reveal non-interbreeding ecologically distinct sympatric populations that differ by less than 4%.…”

“…Our previous work [ 21 – 23 ] used 4% mtCOI divergence as a conservative but arbitrary cut-off to define “species.” Analyses of allozymes [ 38 ] and microsatellites [ 24 ] of sympatric populations show that members of groups (“species”) differing by 4% mtCO1 do not interbreed. Furthermore, microsatellite data [ 24 ] clearly reveal non-interbreeding ecologically distinct sympatric populations that differ by less than 4%. Because we are interested in the process of diversification (divergence), which may not always result in speciation (which is difficult to document when divergent populations are allopatric), we chose here to focus on divergent monophyletic “haplotype lineages” differing from one another by at least three mutations (0.5% divergence).…”

“…From a single individual plant (a male Gurania spinulosa ), we reared four species of Blepharoneura , three of which were observed courting (and were captured in copula ) on the same individual plant. We discovered an even more diverse community along a 1 km transect in Peru: we reared fifteen species of Blepharoneura from flowers of just two host plant species: G. acuminata and G. spinulosa [ 23 ], and more recent work suggests this may have been an underestimate of total diversity [ 24 ]. Such high levels of niche overlap [ 22 – 24 ] are not predicted by classic models of adaptive radiation [ 25 , 26 ].…”