Diverse organisms are associated with obligate microbial mutualists. How such essential symbionts have originated from free-living ancestors is of evolutionary interest. Here we report that, in natural populations of the stinkbug Plautia stali, obligate bacterial mutualists are evolving from environmental bacteria. Of six distinct bacterial lineages associated with insect populations, two are uncultivable with reduced genomes, four are cultivable with non-reduced genomes, one uncultivable symbiont is fixed in temperate populations, and the other uncultivable symbiont coexists with four cultivable symbionts in subtropical populations. Symbiont elimination resulted in host mortality for all symbionts, while re-infection with any of the symbionts restored normal host growth, indicating that all the symbionts are indispensable and almost equivalent functionally. Some aseptic newborns incubated with environmental soils acquired the cultivable symbionts and normal growth was restored, identifying them as environmental Pantoea spp. Our finding uncovers an evolutionary transition from a free-living lifestyle to obligate mutualism that is currently ongoing in nature.
Symbiosis enables organismal evolution and diversification in a variety of ways1,2. Obligate insect–bacterium symbiotic associations are among the most sophisticated forms of symbiotic intimacy, where the host and the symbiont are integrated into a coherent biological entity and are unable to survive without the partnership3,
Symbiont polymorphism in natural host populations
The brown-winged green stinkbug Plautia stali (Hemiptera: Pentatomidae) (Fig. 1a), a notorious pest infesting various fruits and crop plants12, possesses a specific bacterial symbiont in a posterior region of the midgut, where there are numerous sac-like structures, called crypts or caeca, that harbour the bacteria extracellularly (Fig. 1b,c)13,
Mapping of the symbiont types on the Japanese P. stali populations revealed a striking geographic pattern: symbiont A was fixed in the temperate mainlands, and symbionts B, C, D, E and F were coexisting in the subtropical southwestern islands (Fig. 2). In the southwestern islands, the great majority of infections were symbiont B infections (69–100%), and the rates of symbiont C, D, E and F infections were relatively minor, except on Miyako Island, where symbiont C infection was dominant (Fig. 2). Molecular phylogenetic analysis of the insects’ mitochondrial gene sequences revealed that 333 insects representing the mainland and southwestern island populations of P. stali constituted a coherent monophyletic group (Supplementary Fig. 2a), with five distinct clades reflecting the geography: clade I from Ishigaki and allied islands, clade II from Miyako island, clade III from Yakushima, Amami-Oshima and allied islands, clade IV from Okinawa and allied islands, and clade V from the mainlands and Tanegashima Island (Supplementary Fig. 2b). Within each of the clades, except for clade V, the symbiont types did not reflect the host genotypes (Supplementary Fig. 2c–g), suggesting promiscuous host–symbiont associations in the southwestern island populations of P. stali.
All indispensable, some fastidious and others cultivable
All symbionts A–F were vertically transmitted through host generations via egg surface contamination (Supplementary Table 1). When the symbiont was removed by egg surface sterilization, the symbiont-free offspring consistently exhibited severe growth defects and scarcely reached adulthood, revealing that not only the major symbionts A and B but also the minor symbionts C, D, E and F are indispensable for P. stali (Fig. 3a,b and Supplementary Fig. 3a–e). When cultured on Luria-Bertani agar plates, symbionts A and B exhibited no growth, but, strikingly, symbionts C, D, E and F formed numerous colonies (Supplementary Table 2), indicating that the minor symbionts C, D, E and F are cultivable whereas the major symbionts A and B are uncultivable, at least on standard bacterial media. Draft genome sequencing estimated the genome sizes of the major symbionts A and B (2.4–3.9 Mb) to be remarkably smaller than those of the minor symbionts C, D, E and F (4.7–5.5 Mb) (Supplementary Table 3), suggesting reductive genome evolution in the uncultivable symbionts A and B.
Using southwestern island strains of P. stali associated with symbiont B, we experimentally replaced the original symbiont with one of the cultivable symbionts C, D, E or F by means of egg surface sterilization and symbiont supplementation via drinking water (Supplementary Fig. 3f). To our surprise, the symbiont-replaced insects grew normally (Fig. 3c), reached adulthood (although the insects with symbiont F exhibited a lower adult emergence rate because of the inferior establishment of infection; Fig. 3d), and exhibited normal adult phenotypes in agreement with the control insects infected with the original symbiont B (Fig. 3f,g and Supplementary Fig. 3g). The symbiont-replaced females laid eggs normally (Fig. 3e and Supplementary Fig. 3h) and vertically transmitted the replaced symbionts to the offspring (Supplementary Fig. 3i). The same results were obtained in experiments using a mainland strain of P. stali originally associated with symbiont A (Supplementary Fig. 4). These results strongly suggest that, at least under laboratory conditions, symbionts A–F are functionally equivalent to one another and P. stali requires at least one of the symbionts for normal growth and development.
Phylogenetically, symbionts A–F are closely related to the free-living Pantoea species that are commonly found in soils, plants and other environmental sources (Fig. 1e)17. Using the P. dispersa type strain from soil and the P. agglomerans type strain from human wounds deposited in the American Type Culture Collection (Fig. 1e), we performed symbiont-replacing experiments with the P. stali strains originally associated with symbiont B. Strikingly, the environmental Pantoea isolates were able to establish infection and support the growth and survival of P. stali in place of the original symbiont (Fig. 3h and Supplementary Fig. 5a–f), although P. agglomerans exhibited inferior infection establishment and incomplete vertical transmission (Fig. 3h and Supplementary Fig. 5f). By contrast, the unrelated bacteria Escherichia coli, Bacillus subtilis and the Burkholderia symbiont of the bean bug Riptortus pedestris18 were unable to establish infection and support growth and survival of P. stali (Fig. 3i and Supplementary Fig. 5g,h), suggesting that the symbiotic capability may be specific to the Pantoea-allied bacterial strains.
Environmental symbiont pool
We surface-sterilized the eggs of the P. stali strains associated with symbiont B, kept the symbiont-free newborn nymphs with soil samples collected at three P. stali habitats on Ishigaki Island for an initial six days (Supplementary Fig. 6a,b), and subsequently reared them in clean containers. Of 1,005 newborns subjected to the experiment, 920 (91.5%) died as nymphs and 85 (8.5%) reached adulthood. Of the 85 adult insects, 14 (1.4%) were dwarf and yellowish, as is typical of symbiont-deficient insects, while 71 (7.1%) were of normal size and green in colour (Supplementary Table 4 and Supplementary Fig. 6f–o). Molecular phylogenetic analysis of the bacterial genes obtained from their midgut revealed that the majority of the green adults (39 of 71) had established an association with symbiont C, D or E (Fig. 4; Supplementary Figs 1b and 6c–e,g–i). Meanwhile, the remaining 32 green adults were associated with several bacterial lineages (tentatively called bacteria X1–X6) that are placed in the Enterobacteriaceae, and are phylogenetically distinct from symbionts A–F and not found as symbionts in natural populations of P. stali (Fig. 4; Supplementary Figs 1b and 6c–e,j–o). These results indicate that (1) symbionts C, D and E are commonly present in the environment of natural P. stali habitats, (2) newborn nymphs of P. stali are able to acquire and establish infection with symbionts C, D and E from the environment, (3) these ecological aspects may account for the cultivability, low infection frequencies and host–symbiont phylogenetic incongruence observed with symbionts C, D and E, (4) bacteria X1–X6, which are present in the soil environment but not identified as symbionts in natural P. stali populations, can infect sterilized newborns of P. stali and support their growth and survival, and (5) presumably, these ‘potential’ symbiotic bacteria in the environment may comprise the source for evolution of new obligate symbiotic bacteria like symbionts A–F. Although speculative, the absence of bacteria X1–X6 in natural P. stali populations might be attributable to their inferior infection/proliferation in the midgut of P. stali, their inefficient vertical transmission to the offspring of P. stali, their lower competitiveness to the resident symbionts A–F within P. stali, or a combination of these factors.
Interspecific symbiont sharing
In the southwestern islands of Japan where symbionts B, C, D, E and F coexist in natural P. stali populations, we found that some stinkbug species are associated with symbiotic bacteria very closely related to symbionts C, D and E: for example, Solenosthedium chinense with symbiont C; Lampromicra miyakona with symbionts C or D; Carbula crassiventris with symbiont D; Axiagastus rosmarus with symbionts C, D or E (Fig. 5a). We isolated and cultivated symbiont C from S. chinense, symbiont D from L. miyakona and symbiont E from A. rosmarus (Supplementary Table 2) and performed symbiont-replacing experiments with P. stali strains associated with symbiont B. These heterospecific symbionts successfully established infection, supported normal growth, survival and reproduction, and were vertically transmitted to the next generation of P. stali (Fig. 5b and Supplementary Fig. 7). These results suggest that, although normally transmitted vertically via egg surface contamination, symbionts C, D and E may occasionally be acquired from the environment and also transferred from other stinkbug species horizontally. Through these pathways, P. stali and other stinkbugs may share the common environmental symbiont pool ecologically. In this context, it is notable that Pantoea-allied gut symbiotic bacteria have been identified from diverse stinkbug lineages14,19,
In this study, we have described an unprecedented snapshot of currently ongoing evolution from environmental free-living bacteria to obligate uncultivable bacterial mutualists in natural insect populations. Strikingly, in Japanese populations of P. stali, all evolutionary stages coexist (from free-living through facultative to obligate associations): environmental bacteria that are potentially capable of symbiosis (bacteria X1–X6); symbiotic bacteria that are capable of both symbiosis and free-life (symbionts C–F); a symbiotic bacterium that is obligatorily symbiotic, incapable of free life and uncultivable, but not yet fixed in host populations (symbiont B); and a symbiotic bacterium that is obligatorily symbiotic, incapable of free life and uncultivable and completely fixed in host populations (symbiont A). We expect that, in field monitoring of P. stali–bacteria associations, we would be able to observe the process of symbiotic evolution in natural populations, as in the evolution of Darwin's finches in the Galapagos Islands22. The P. stali–bacteria associations will also provide an ideal system for experimental evolutionary studies, in which less specialized or potential symbiotic bacteria can be selected for or against to be infectious to, stable in and beneficial for the insect host.
Of the P. stali-associated symbiotic bacteria, the relationship between symbiont A and symbiont C is of particular interest in that (1) symbiont A is phylogenetically very close to symbiont C or P. dispersa (Fig. 1e), (2) symbiont A exhibits a conspicuously elongated branch in the phylogeny (Fig. 1e) and (3) the genome of symbiont A seems remarkably smaller than the genome of symbiont C (Supplementary Table 3). These patterns suggest the possibility that, although speculative, symbiont A derived from an environmental bacterium allied to symbiont C or P. dispersa has experienced reductive genome evolution through sustained host–symbiont association5,
How symbionts A–F, respectively, contribute to host P. stali deserves future studies. In other stinkbugs, genomic and nutritional studies have revealed the provisioning of essential amino acids and/or vitamins by their gut symbiotic bacteria8,9,23. Considering that many Pantoea species and strains are genetically tractable24, molecular and physiological mechanisms underlying P. stali gut symbiosis would be uncovered by making use of the cultivable Pantoea-allied symbionts C–F. In the Japanese Archipelago, symbiont A is fixed in P. stali populations across the temperate mainlands, whereas symbiont B is dominant in P. stali populations across the subtropical islands (Fig. 2). The possibility that the different obligate gut symbionts may be involved in the host's adaptation to different climatic conditions25,26 or different host plants27,28 also deserves future studies.
Conventionally, highly specialized obligate symbiotic bacteria have been viewed as microbial entities distinct from environmental free-living bacteria, but the apparent gap between them has been blurred in both ecological and evolutionary contexts29,30. For a better understanding of insect–bacterium and other forms of symbiotic associations, an evolutionary continuum ranging from free-living through facultative to obligate associations should be considered properly. Throughout the world, numerous organisms are infected with a myriad of microorganisms, potentially leading to the establishment of obligate and mutualistic symbiotic associations. It is plausible that such symbiotic associations are coming into being and evolving at any time, everywhere.
Insect sampling and rearing
Field-collected samples of P. stali and other stinkbugs are listed in Supplementary Table 5. Adult insects of P. stali were mainly collected by light trapping or sweeping mulberry trees Morus australis (Moraceae). Some insects were collected from other plants including Pittosporum tobira (Pittosporaceae), Diplocyclos palmatus (Cucurbitaceae), Melothria liukiuensis (Cucurbitaceae), Cerasus jamasakura (Rosaceae), Sambucus formosana (Caprifoliaceae) and Rhus succedanea (Anacardiaceae). From some of the field-collected insects, isofemale lines were established and maintained with dry soybean seeds (Glycine max), raw peanuts (Arachis hypogaea) and water containing 0.05% ascorbic acid at 25 °C in sterilized plastic Petri dishes under a long-day regimen (16 h light and 8 h dark)31.
DNA cloning and sequencing
The symbiotic organ (crypt-bearing midgut fourth section) was dissected from each adult insect, washed well with sterilized phosphate-buffered saline (PBS: 137 mM NaCl, 8.1 mM Na2HPO4, 2.7 mM KCl and 1.5 mM KH2PO4) and cut into halves. One half was subjected to DNA extraction using QIAamp DNA Mini Kit (QIAGEN), and the other half was tested for cultivability of the symbiont. For nymphal insects, DNA was extracted from the whole body. From each DNA sample, bacterial 16S rRNA and groEL genes and the insect mitochondrial 16S rRNA gene were amplified by PCR, cloned, genotyped and sequenced as described28,32.
Molecular phylogenetic analysis
We designed primer sets to specifically target the 16S rRNA gene or groEL gene of the symbiotic and other bacteria as follows: [5′-TTG TGC CCT TGA GGC GTA A-3′] and [5′-GAA TCA CAA AGT GGT AAG CGT-3′] for a 650 bp region of the 16S rRNA gene of symbiont A; [5′-GTA CTG GCG CAG TCT ATY GTT AG-3′] and [5′-TTC GCG GCC AGA GAT AGC AGA-3′] for a 760 bp region of the groEL gene of symbiont B; [5′-GAG CTG GAA GAC AAG TTC GAG-3′] and [5′-ATG AAT GGG CTT TCC ARC TCC-3′] for a 479 bp region of the groEL gene of symbiont C and P. dispersa; [5′-CTG CAG GTG ACG GTA CG-3′] and [5′-AGG ATG TAC GGT GCA TCC AGT-3′] for a 412 bp region of the groEL gene of symbiont D; [5′-TAT CCT GTT GGC TGA CAA GAA G-3′] and [5′-CTG ACC TTT CAG RCC AGA AGC A-3′] for a 645 bp region of the groEL gene of symbiont E; [5′-TTT GAT CGC GGC TAC CTG TCA-3′] and [5′-GCC CTG AAT GGT GTT CTC TTC A-3′] for a 450 bp region of the groEL gene of symbiont F; [5′-AGC CGG AAA CTG GCG CAG TA-3′] and [5′-ACG CAG GTC AGC CAG TTT AGA-3′] for a 675 bp region of the groEL gene of E. coli; [5′-AGG AGG TCT TGT AAA CAT GGC-3′] and [5′-CTT TAG AAG CGC GAC CCA ATT-3′] for a 977 bp region of the groEL gene of B. subtilis; [5′-TTT TGG ACA ATG GGG GCA AC-3′] and [5′-GCT CTT GCG TAG CAA CTA AG-3′] for a 787 bp region of the 16S rRNA gene of Burkholderia sp.; [5′-CGA TGA AAC CGT GGG CCA GT-3′] and [5′-CAG ATC TTC CAG AGC CGC TTT C-3′] for a 490 bp region of the groEL gene of P. agglomerans (Supplementary Fig. 8). Diagnostic PCR detection of these bacteria was performed for each insect DNA sample.
Each dissected symbiotic organ was ground in sterilized PBS, spread on Luria-Bertani agar plates and incubated at 25 °C for several days. If colonies were formed, they were subjected to DNA extraction, PCR and sequencing of the 16S rRNA gene. The bacterial isolates were stored as glycerol stocks at –80 °C.
Symbiont genome analysis
DNA libraries were constructed according to the Roche Protocol with the GS FLX Titanium Rapid library kit. The library was sequenced using a GS FLX Titanium Sequencing Kit XLR70 on the Roche 454 GS FLX Titanium sequencing platform. Sequence reads were assembled with the GS De Novo Assembler v2.6 (454 Life Sciences, Roche). To remove host-derived contigs from the total contig pools, each contig of symbionts A and B, the total DNA of which was extracted with host tissues, was divided into lengths of 1,000 nucleotides overlapping by 500 nucleotides. Using these DNA fragments as query, BLASTX similarity searches were performed against the custom database containing the proteomes of 12 bacterial species and 7 insects. Contigs whose divided fragments were significantly more similar to bacterial genes than to insect genes were judged as contigs of the symbiotic bacteria8. Gene predictions were conducted by Glimmer v3.0235. Small fragmented open reading frames (ORFs) interrupted by internal frame shifts, premature stop codons, large insertions and deletions were designated pseudogenes.
From each isofemale line of P. stali, two egg masses were randomly sampled and assigned to either the control treatment (in which the egg mass was soaked in sterilized water for 40 min) or the sterilized treatment (in which the egg mass was soaked in 70% ethanol for 10 min and subsequently in 4% formaldehyde for 30 min and finally rinsed with 70% ethanol). After air-drying, each egg mass was kept in a sterilized plastic Petri dish. The hatchlings were supplied with sufficient food and water, and subjected to fitness measurements.
Symbiont replacement by cultured bacteria
Isofemale lines of P. stali originating from symbiont B-infected adult insects collected at Ishigaki Island, Japan, or those originating from symbiont A-infected adult insects collected at Tsukuba, Japan, were used. The original symbiont was experimentally replaced by the following bacterial isolates: symbiont C from P. stali (insect sample ID Ps-ISGKf53; strain accession no. JCM30570); symbiont D from P. stali (Ps-ISGKm56; JCM30571); symbiont E from P. stali (Ps-ISGKf70; JCM30572); symbiont F from P. stali (Ps-KMJMf21, JCM30573); symbiont C from S. chinense (Sc-ISGK; JCM30574); symbiont D from L. miyakona (Lm-IROM; JCM30575); symbiont E from A. rosmarus (Ar-MYKJ; JCM30576); P. dispersa type strain (ATCC14589); P. agglomerans type strain (ATCC27155); E. coli strain K12 (ATCC10798); B. subtilis subsp. spizizenii (ATCC6633); and Burkholderia sp. RPE6718. From each isofemale line, three to five egg masses were randomly sampled, one of which was assigned to the control treatment in which the egg mass was soaked in sterilized water for 40 min. The other egg masses were assigned to the symbiont-replacing treatments. After surface sterilization as described above, each of the egg masses was individually placed in a sterilized Petri dish. The hatchlings in the control treatment were supplied with 3.5 ml sterilized water, and the hatchlings in the symbiont-replacing treatments were provided with 3.5 ml cultured bacterial suspension (107–108 c.f.u. per ml) using the setting shown in Supplementary Fig. 3f. In both the control and symbiont-replacing treatments, the nymphs were provided with food from the fourth day after hatching, when they moulted to the second instar. Note that first instar nymphs of P. stali require only water but no food to moult to the second instar. On the seventh day after hatching, the nymphs from each of the experimental egg masses were transferred to a new sterilized Petri dish and supplied with sufficient food and water. Fitness measurements were performed until all the nymphs either became adult or died. The adult insects were either subjected to DNA extraction and diagnostic PCR detection of the symbiotic bacteria, or allowed to mate and reproduce. Eggs laid by the adult females were counted and their offspring were subjected to diagnostic PCR detection of the vertically transmitted symbiotic bacteria at the second instar.
Symbiont acquisition from environmental soil
Soil samples were collected at three localities: Inoda, Nosoko and Yarabu in Ishigaki Island, Japan (Supplementary Fig. 6b). All sites were forest edges where P. stali was commonly found on the host trees. Soil samples were brought to the laboratory, sieved at a mesh size of 2 mm and stored at 4 °C until use. In total, 106 egg masses were sampled from seven isofemale lines originating from symbiont B-infected adult insects collected at Ishigaki Island. The egg masses were surface-sterilized as described above and randomly assigned to the four experimental treatments: (1) control treatment, (2) Inoda soil treatment, (3) Nosoko soil treatment and (4) Yarabu soil treatment. In the control treatment, hatchlings from a sterilized egg mass were kept in a sterilized Petri dish with a piece of cotton soaked with distilled water as shown in Supplementary Fig. 3f, whereas in the soil treatments, hatchlings from a sterilized egg mass were kept in a sterilized Petri dish with about 6 g of soil soaked with distilled water as shown in Supplementary Fig. 6a. In all treatments, the nymphs were provided with food from the fourth day after hatching, transferred to a new sterilized Petri dish with sufficient food and water on the seventh day after hatching and reared until all the insects either became adult or died.
The following fitness parameters were recorded for the experimental insects: adult emergence rate, nymphal period from egg hatching to adult moulting, fecundity in terms of number of eggs laid by a female within 10 days after the first oviposition, thorax width of adult males, and thorax width of adult females. Adult emergence rate was calculated for each rearing container, and the number of replicates corresponded to the number of rearing containers used in the experiment. The other fitness parameters were recorded for each individual insect and thus the number of replicates reflected the number of insects examined in the experiment. Nymphal period and thorax width were recorded for all adult insects, whereas fecundity was recorded for selected adult females.
The nucleotide sequences of bacterial 16S rRNA and groEL genes, and insect mitochondrial 16S rRNA gene reported in this study were deposited in the DNA Data Bank of Japan under sequence accession numbers LC007115–LC007454 and LC007457–LC007855. The DNA Data Bank of Japan accession numbers of draft genome sequences reported in this study are BBNZ01000001-135, BBOA01000001-35, BBOB01000001-220, BBOC01000001-280, BBOD01000001-165 and BBOE01000001-141. Some bacterial isolates used for symbiont-replacing experiments were deposited at the Japan Collection of Microorganisms (www.jcm.riken.jp) under strain accession numbers JCM30570 to JCM30576.
The authors thank M. Baba, N. Baba, Y. G. Baba, D. Haraguchi, H. Hirayama, M. Hironaka, S. Kada, N. Kaiwa, Y. Kikuchi, K. Kizaki, S. Kudo, T. Makino, M. Moriyama, N. Nagata, S. Ohno, T. Ohtani, M. Ono, M. Sakakibara, H. Toju, K. Tsuji, N. Tsurusaki, T. Uesato, R. Ukuda, H. Watanabe and T. Yamaguchi for insect samples; J. Makino, N. Tanifuji, U. Asaga and W. Kikuchi for technical assistance; J. P. McCutcheon for comments on the manuscript; and Y. Nakajima and Y. Kikuchi for logistic and technical support. This study was supported by a JSPS KAKENHI grant (25221107) to T.F. and by the University of the Ryukyus Foundation to T.H., and Y.I. was supported by a JSPS Fellowship for Young Scientists.
Supplementary Figures 1–8 and Tables 1–5.
About this article
Aseptic rearing procedure for the stinkbug Plautia stali (Hemiptera: Pentatomidae) by sterilizing food-derived bacterial contaminants
Applied Entomology and Zoology (2017)