Uneven transfer amount from bacteria to insects
Although the five available insect genomes were analyzed for HGT, significant HGT events were detected only in the silkworm. Three reasons may explain this observation. The first is utilization of incomplete sequence data in this study. In general, before the assembly of a genome sequence, the reads are checked to get rid of contaminant fraction from symbiont, parasite and pathogen. As a result, HTG sequences coming from other organisms may be deleted as contaminants. The second is that the power of detection methods used in this study is not high. Similarity search and phylogenetic analyses and other information were combined as a detection pipeline to reveal HGT event; this reduces the ratio of false positives. However, the complex steps and strict standards in the pipeline may also reduce the number of the detected candidates. The third is that the contrast in HTG numbers between the silkworm and other four insects may be true.
Wolbachia species are endosymbionts infecting 20% of arthropods on Earth [66, 67], and the genetic fractions of this prokaryote were detected in the genomes of numerous insects [22, 23, 68, 69]. This suggests that these insects have ability to accept foreign genetic materials. Additionally, the D. melanogaster genome was first sequenced among the five insects, and its annotation is more complete and detailed. However, the trail of HTG was not found in D. melanogaster based on its sequence data. Consequently, incomplete data is not a reasonable explanation. Various types and large amount of Lepidopteran HTGs do exist. In addition, we found 79 suspected sequences in Ap. mellifera, which are considered as contaminants. Strikingly, all previously reported silkworm transferred genes were recovered in this study [18–20, 27], especially, a recent and independent study on the silkworm HGT also confirmed the methods and results in our study . Thus, the detection pipeline we used appears to be powerful and should not miss real HTGs. In conclusion, HGT events may be distributed unevenly at least in four major insect orders of the five insects. Silkworm (Lepidoptera) is a distinct case in which HTGs are common and functional. The disparity of the transfer number and ratio of HTGs in a certain organism category has been shown previously in an HGT study on fungi .
The proportion (0.15%) of the transferred genes in silkworm genome is comparable with the average level (0.12%) in fungi . In chromalveolates, a group of protista, 16 types of bacteria transferred genes were detected . However, as simple eukaryotes, fungi and protista are thought to accept foreign genetic sequences into their genomes more easily. In aphid, 6 types of 12 genes including 3 pseudogenes were detected as HTGs and its proportion ranges from 0.11% to 0.03% (0.08% to 0.03% when deleting pseudogenes) [26, 71]. Thus, the amount of Lepidopteran HTGs is relative large. Whether some ancient Lepidopteran insects have an unusual ability of acquiring and fixing foreign genetic materials is unclear. A previous study on prokaryotes suggested that the proportion of distant HTGs is correlated to the genome sizes of donors and recipients . Therefore, a large proportion of HTGs in Lepidopteran insects may be, in part, attributed to the genome structure and component (e.g., genome size, transposable element, recombination rate, metabolic network and so on) of these recipients. However, difference in selection pressure among insects due to their surrounding niches is another determinant and may result in uneven transfer amount among bacteria and insects.
The evolutionary characteristics of Lepidopteran HTGs
There are common features of HGT in prokaryotes and eukaryotes. In prokaryotes, the sequence composition of anciently transferred genes are often ameliorated to the host genome . As expected, GC contents of the transferred genes in silkworm display a more centralized distribution (mean: 43.83%, s.d.: 4.82%) compared with that of the predicted bacterial donors (mean: 49.23%, s.d.: 11.27%; Additional file 8). This also indicates that most of these transferred genes have been integrated into recipient genome for a long period of time, which is consistent with the EST search results in Lepidoptera (Figure 3). Three of the 14 types of Lepidopteran HTGs (β-fructofuranosidase, glycerophosphoryl diester phosphodiesterase and NAD-dependent epimerase/dehydratase) contain multiple genes. The glucose-1-phosphatase consists of 6 tandem arranged genes (Table 2). However, these gene duplication events were not revealed in a recent study . In aphid, the transferred gene of rare lipoprotein A (RlpA) has 5 duplicated genes which are also tandem arranged, and another gene type, LD-carboxypeptidase_1 (LdcA), includes 2 genes with one inactive . At least more than a quarter of the detected HTGs were duplicated after HGT events, either in Lepidoptera or in aphid. This tendency is consistent with the findings in prokaryotes that the transferred genes are more frequently duplicated than endogenetic genes in hosts . We also found that some detected HTGs (BGIBMGA002521, BGIBMGA007766 type, BGIBMGA008709) harbour respective homologs which are vertically transferred genes in the silkworm and other insect genomes; that is, these HTGs and their homologs belong to the same protein family. In a recent study, the significant contribution of HGT to the expansion of protein families in bacteria was revealed . Thus, HGT events also affect the evolution of protein families in Lepidopteran insects, more or less.
There are some special aspects for the HTGs detected in this study. Potential prokaryotic donors of insects include symbionts, parasites, pathogens and bacteria in diet and surrounding environments. Previous studies on HGT between prokaryotes and insects mainly focused on the insect endosymbiont bacteria . In this paper, we also found that one gene (BGIBMGA009498) might be transferred from Wolbachia bacterium. However, the experimental evidence indicated that extant silkworm may be not infected by Wolbachia species . Similar results were found in Aedes aegypti and two filarial nematode species [24, 76]. At least 7 types of HTGs may be introduced from entomopathogenic bacteria (pathogenic bacteria) and another donor is an endophytic bacterium (bacteria in food) (Table 3). Thus, donor pattern of Lepidopteran HTGs is multifarious, which is different from that in aphid and other insects. A recent HGT event can be detected more easily than ancient ones, because sequence similarity between donor and the transferred sequence will decrease and base composition of transferred sequence will ameliorate to the recipient genome after the fixation . In aphid, RlpA gene was transferred from a relatively ancient HGT event about 50-70 MYA ago . In HGT studies between insects and Wolbachia, the majority of genes were transferred from endosymbiont to hosts recently. Based on phylogenetic topologies and homolog distribution (Figure 3), we found that most of Lepidopteran HTGs might be integrated into the hosts at least before the radiation of Ditrysia or Apoditrysia group, about 100 MYA ago . This indicates that the method we used is effective in revealing ancient HGT events. Additionally, these detected Lepidopteran HTGs, as a group of special-origin genetic fragments, can be used in the phylogenetic reconstruction of Lepidopteran insects, especially for Ditrysia insects among which the phylogenetic relationships are not clear in detail .
The biological significance of Lepidopteran HTGs in the evolution of hosts
A question may arise: why do the transfer times of most detected HGT events in Lepidoptera fall within a relatively narrow evolutionary period? Lepidopteran insects are a relative young biological group in geology history compared with other insect orders. Furthermore, Lepidoptera are the second largest order in insecta and the largest group in plant-feeding insects. A generally accepted opinion holds that the prosperity of Lepidopteran insects is associated with the diversification of angiosperms on Earth in the late Cretaceous period which is just about 100 MYA ago . At that time, developing angiosperms provide rich foods and living environments for some ancient Lepidopteran insects (may be the ancestor of Ditrysia insects). Thus, the genetic changes that facilitated hosts to adapt to the new niches predominated by angiosperm would be fixed in the ancient Lepidopteran lineages. Biological traits related to nutrition, reproduction, defense and immunization, are major targets of natural selection. Almost all of the detected HTGs are functional enzymes except for BGIBMGA000070 which may be a storage protein. Furthermore, previous studies revealed that one HTG BGIBMGA007146 in the silkworm participated in degradation, modification and combination of the toxins , and another HTG BGIBMGA011204 improved the metabolic pathway to get out of the toxic target site(s) . Again, the silkworm HTG BGIBMGA005696 is involved in replacement of the targeted enzyme with a resistant one . Thus, the majority of Lepidopteran HTGs might perform physiological functions in nutritional metabolism and detoxification. Detoxification is related to the nutritional metabolism because toxins in diet and toxins generated in normal endogenetic metabolisms can obstruct and reduce the ingestion and digestion activities of consumers. In a long interaction history between insects and their dietary plants, phytophagous insects might have developed some effective strategies to protect the efficiency of nutritional metabolism from the detriments produced by plants. The HTGs from bacteria may have contributed novel functions for Lepidopteran hosts to adapt to various diets and niches.
It is interesting that a half of the detected HTGs are pathogen-origin, while previous studies indicated that bacteria of endosymbionts and parasites may be the primary donor group of HGT. It is well known that obligate symbiotic and parasitic bacteria are often degenerated in some physiological metabolisms and phenotypic traits [78–80]. Thus, potential genetic variations they could offer for recipients may be not plentiful and effective to increase host's adaptability in complex niches, especially when surrounding environments shifted [81, 82]. This is probably a reason that most of symbiont-origin HTGs in insects and nematodes are nonfunctional or inactivated [23, 25, 26]. In contrast, pathogenic bacteria in this study are more complex in ecological niche and biological functions compared with obligated symbiotic bacteria [56, 58–60]. Previous studies revealed several HGT events from non-endosymbiosis organisms to multicellular recipients [9, 16, 26]. However, the mechanism of foreign sequences transferred into the recipient germline from non-endosymbiosis organisms is unclear. Thus, how these pathogen genes integrated into the Lepidopteran insects remains to be elucidated.
Whether HGT from other organisms to multicellular eukaryotes has biological significance in the evolutionary process of hosts is unknown. Two factors may affect the evolutionary significance of the transferred genes: their persistence in host genome and their integration in host biology . On the basis of previous individual cases, some transferred genes certainly perform functions in the host. However, in this study, we found a relative large group of ancient transferred genes in Lepidopteran insects, and these genes are predicted to have biological functions since they were integrated into ancient Lepidopteran genomes. Furthermore, the integration of these HTGs into Lepidoptera at least corresponds to the expansion of angiosperm. Thus, it is most likely that most of these HTGs facilitated Lepidoptera to adapt to the evolution of their plant hosts. In short, our results provide new evidence to support for exogenic variations significantly contributing to the evolution of metazoan organisms.