Localization and transmission route of Coriobacterium glomerans, the endosymbiont of pyrrhocorid bugs


  • Editor: Christoph Tebbe

Correspondence: Martin Kaltenpoth, Zoological Institute, University of Regensburg, 93053 Regensburg, Germany. Tel.: +49 941 943 3057; fax: +49 941 943 3325; e-mail: martin.kaltenpoth@biologie.uni-regensburg.de


Endosymbiotic gut bacteria play an essential role in the nutrition of many insects. Most of the nutritional interactions investigated so far involve gammaproteobacterial symbionts, whereas other groups have received comparatively little attention. Here, we report on the localization and the transmission route of the specific actinobacterial symbiont Coriobacterium glomerans from the gut of the red firebug, Pyrrhocoris apterus (Hemiptera: Pyrrhocoridae). The symbionts were detected by diagnostic PCRs and FISH in the midgut section M3, in the rectum and in feces of the bugs as well as in the hemolymph of some females. Furthermore, adult female bugs apply the symbionts to the surface of the eggs during oviposition, from where they are later taken up by the hatchlings. Surface sterilization of egg clutches generated aposymbiotic insects and thereby confirmed the vertical transmission route via the egg surface. However, symbionts were readily acquired horizontally when the nymphs were reared in the presence of symbiont-containing eggshells, feces, or adult bugs. Using diagnostic PCRs and partial sequencing of the 16S rRNA gene, closely related bacterial symbionts were detected in the cotton stainer bug Dysdercus fasciatus (Hemiptera: Pyrrhocoridae), suggesting that the symbiosis with Actinobacteria may be widespread among pyrrhocorid bugs.


Many insects are intimately associated with microorganisms that reside in the digestive tract or in specialized cells or organs of the host (Buchner, 1965; Bourtzis & Miller, 2003). While some facultative symbionts (e.g. Wolbachia) can cause reproductive alterations in the insect host or confer other negative effects, most obligate and several facultative symbioses are of mutualistic nature and significantly enhance the fitness of the host (Douglas, 1998; Bourtzis & Miller, 2003). In the majority of these cases, the symbionts produce essential nutrients that the insects can neither synthesize themselves nor obtain in sufficient quantities from the diet (Douglas, 1998, 2006; Shigenobu et al., 2000; Gil et al., 2003; Ishikawa, 2003; Zientz et al., 2004). However, several cases of defensive associations have recently been described in which the symbionts confer protection to their hosts against pathogens or parasitoids (Currie et al., 1999; Kellner, 2002; Oliver et al., 2003; Kaltenpoth et al., 2005; Scarborough et al., 2005; Scott et al., 2008).

Although environmental uptake of symbionts in each host generation has recently been reported for a mutualistic insect symbiont (Kikuchi et al., 2007), vertical transmission from mother to offspring appears to be the predominant mode of symbiont transfer in insects (e.g. Buchner, 1965; Baumann & Moran, 1997; Bourtzis & Miller, 2003). However, the routes of vertical transmission differ among symbiotic systems. The intracellular primary symbionts of aphids, carpenter ants, weevils, and several other insect taxa are generally unable to survive outside of the hosts' cells and have to be transmitted transovarially to the eggs (Buchner, 1965; Schroder et al., 1996; Douglas, 1998; Sauer et al., 2002; Nardon, 2006) or via the milk glands to the developing larvae, as it seems to be the case in the pupiparous tse-tse flies (Buchner, 1965; Aksoy et al., 1997). Many extracellular symbionts, however, are transmitted by posthatch mechanisms that require the symbionts to survive outside the hosts for a part of their life cycle (e.g. Buchner, 1965; Fukatsu & Hosokawa, 2002; Hosokawa et al., 2005; Kaltenpoth et al., 2005; Prado et al., 2006).

An especially large diversity of posthatch vertical transmission mechanisms has been described for gut endosymbionts of Hemipteran insects. While the nymphs of some species acquire the symbionts by the ingestion of excretes from adult individuals (Baines, 1956; Buchner, 1965; Beard et al., 2002), others take up symbiotic bacteria by probing secretions smeared onto the egg surface by the mother during oviposition (Rosenkranz, 1939; Buchner, 1965; Prado et al., 2006). Plataspid stinkbugs have evolved an especially interesting transmission mechanism involving the formation and deposition of symbiont-containing capsules that are later probed by the hatchlings for the acquisition of the symbionts (Schneider, 1940; Fukatsu & Hosokawa, 2002; Hosokawa et al., 2005).

The Pyrrhocoridae are a family of phytophagous landbugs comprising about 300 species, many of which exhibit aposematic colorations. Although the red firebug (Pyrrhocoris apterus) constitutes an extensively studied model species, little is known about the bacterial symbionts of this species and of the family Pyrrhocoridae in general (Socha, 1993). However, a high-GC gram-positive bacterium (Coriobacterium glomerans) has been described as an extracellular gut endosymbiont of P. apterus (Haas & König, 1987, 1988). The symbiont cells form long chains (up to 150 μm) in the midgut of the bugs where they are assumed to aid in the digestion of the diet (Haas & König, 1987), which in Central Europe mainly consists of dry linden seeds (Tilia cordata and Tilia platyphyllos) (Tischler, 1959; Socha, 1993). Because C. glomerans could only be cultivated under anaerobic conditions and was not detected on eggs of P. apterus, Haas & König (1987, 1988) assumed that the nymphs acquire the symbionts by direct ingestion of excretes from adult individuals.

In the present study, we investigated the localization of C. glomerans in red firebugs using molecular techniques (PCR and FISH), and we elucidated the transmission route of the symbionts. Sequencing of parts of the 16S rRNA gene allowed the phylogenetic placement of the symbionts of P. apterus and another pyrrhocorid species, Dysdercus fasciatus, within the Actinobacteria.

Materials and methods

Specimens and rearing conditions

Adult specimens of P. apterus (Hemiptera: Pyrrhocoridae) were collected beneath linden trees (T. cordata and T. platyphyllos) on the campus of the University of Regensburg, Germany. The bugs were kept in plastic containers (20 × 20 × 6 cm) at a constant temperature of 28 °C and long light regimes (16 h/8 h light/dark cycles) to keep the animals in the reproductive stage (Saunders, 1983, 1987). They were provided ad libitum with water and crushed dry linden seeds (T. cordata and T. platyphyllos). Adult D. fasciatus (Hemiptera: Pyrrhocoridae) were obtained from a laboratory culture at the University of Wuerzburg, Germany, which had originally been established with specimens collected in Côte d'Ivoire, West Africa.

DNA extraction and PCR screen for the localization of the symbionts

Eight adult male and female P. apterus individuals, respectively, were killed by freezing for 1 h at −20 °C. The abdomen was incised on both sides, and 0.5–2.0 μL hemolymph was collected with a pipette. The digestive tract and the testes or ovaries, respectively, were collected under distilled H2O (dH2O). The hemolymph samples were transferred into 150 μL tissue and cell lysis solution (MasterPure DNA Purification Kit, Epicentre Technologies) and incubated for 20 min at 37 °C after the addition of 1 μL lysozyme (50 μg μL−1). The gut, testes, and ovaries were submerged in liquid nitrogen and crushed with a sterile pestle. Subsequently, the DNA was extracted from all samples using the MasterPure DNA Purification Kit (Epicentre Technologies) according to the manufacturer's instructions. Two different methods were used to isolate C. glomerans DNA from the eggs of P. apterus: (1) egg clutches were harvested after oviposition, and each egg clutch was crushed under liquid nitrogen and the DNA was extracted as described above (crushed eggs); and (2) 5–20 eggs were transferred into an Eppendorf tube containing 40 μL of buffer DNA-A/B [100 mM Tris/HCl (pH 7.5), 30 mM NaCl, 20 mM EDTA, and 1% SDS (w/v)], then the bacteria were detached from the egg surface by a short ultrasonic treatment (1 min) and 5 s of vortexing. Then the eggs were removed and discarded, and the DNA was extracted from the egg washes as described for the hemolymph samples, including the lysozyme pretreatment (egg washes).

For the detection of C. glomerans, the specific primer pair Cor-2F/Cor-1R was designed on the basis of 16S rRNA gene sequences obtained from the GenBank database (C. glomerans sequence: accession number X79048; Rainey et al., 1994) (Table 1), and the specificity was checked using the probe match option of the Ribosomal Database Project (Larsen et al., 1993; Maidak et al., 2001). To control for failures in the DNA extraction protocol (e.g. due to low DNA quality or quantity), positive control PCRs were set up with primers targeting the 18S rRNA gene of P. apterus (Table 1). PCR amplifications were performed on a Biometra® T-Gradient Thermocycler in total reaction volumes of 12.5 μL containing 1 μL of template, 1 × PCR buffer (20 mM Tris-HCl, 16 mM (NH4)2SO4, and 0.01% Tween 20), 2.5 mM MgCl2, 240 μM dNTPs, 0.8 μM of each primer, and 0.5 U of Taq DNA polymerase (Peqlab Biotechnologie GmbH, Erlangen, Germany). Cycle parameters were as follows: 3 min at 94 °C, followed by 32 cycles of 94 °C for 40 s, 68/66 °C for 1 min (Cor/Pyr primers, respectively), and 72 °C for 1 min, and a final extension time of 4 min at 72 °C.

Table 1.   Primers and probes used for the specific detection and sequencing of Coriobacterium glomerans
PrimerPrimer sequence (5′–3′)Fwd./rev.5′ mod.TargetUse
Cor-2FGGTAGCCGGGTTGAGAGACCFwd. Coriobacteriaceae 16S rRNA geneSpecific Cor-PCR
Cor-1R*ACCCTCCCMTACCGGACCCRev. Coriobacterium/Collinsella 16S rRNA geneSpecific Cor-PCR
Cor-3FGAACGGCACCCACCCTCGFwd. Coriobacterium 16S rRNA geneCor. sequencing
fD1AGAGTTTGATCCTGGCTCAGFwd. Eubacteria 16S rRNA geneCor. sequencing
rP2ACGGCTACCTTGTTACGACTTRev. Eubacteria 16S rRNA geneCor. sequencing
Pyr18S_2FGGGAGGTAGTGACAAAAAATAACGFwd. Pyrrhocoridae 18S rRNA genePyr. control-PCR
Pyr18S_4RGTTAGAACTAGGGCGGTATCTGRev. Pyrrhocoridae 18S rRNA genePyr. control-PCR
Cor653-FITCCCCTCCCMTACCGGACCCRev.FITCCollinsella/Coriobacterium 16S rRNA geneCor-FISH
Cor653-Cy3CCCTCCCMTACCGGACCCRev.Cy3Collinsella/Coriobacterium 16S rRNA geneCor-FISH
EUB338-Cy3§GCTGCCTCCCGTAGGAGTRev.Cy3Eubacteria 16S rRNA geneControl FISH (Eubacteria)
SPT177-Cy3CACCAACCATGCGATCGGTARev.Cy3Ca. Streptomyces philanthi 16S rRNA geneNegative control FISH

Almost complete 16S rRNA gene sequences of the Coriobacterium symbionts of three P. apterus and D. fasciatus individuals, respectively, were obtained for phylogenetic analyses. The DNA of adult bugs was extracted as described above, and two overlapping fragments of the Coriobacterium 16S rRNA gene were selectively amplified using the primers fD1/Cor-1R (633 bp) and Cor-3F/rP2 (1426 bp) (for fD1 and rP2 see Weisburg et al., 1991). PCR conditions were the same as described above, with annealing temperatures of 66 °C (fD1/Cor-1R) and 61 °C (Cor-3F/rP2), respectively. The PCR products were purified using the peqGOLD MicroSpin Cycle-Pure Kit (Peqlab Biotechnologie GmbH) and sequenced commercially with primers fD1 and rP2, respectively (Seqlab Sequence Laboratories, Goettingen, Germany). Sequences were aligned to reference sequences obtained from the GenBank database using the clustalw algorithm implemented in bioedit The alignment was checked by eye and then imported into clustalx 1.83 (Thompson et al., 1997) for the construction of phylogenetic trees by neighbor joining and subsequent statistical evaluation by bootstrapping (100 replicates).


Gut samples, eggs and feces of P. apterus were tested for the presence of C. glomerans using FISH. For specific localization, the digestive tracts were separated into the sections M1, M2, M3, M4, ileum (including the Malpighian tubules) and rectum, respectively (see Haas & König, 1987), and homogenized in dH2O. Eggs were washed in DNA-A/B and treated with ultrasound and vortexing as described above. Hemolymph was extracted from dissected animals as described above or by severing a leg of an anesthetized individual and collecting the drop of hemolymph appearing at the injured leg. All samples were heat-fixated to eight-field microscope slides and dehydrated using an increasing ethanol series. After 20 min incubation with 30 mg mL−1 lysozyme in phosphate-buffered saline at 30 °C, hybridization and washing was carried out as described earlier (Grimm et al., 1998; Kaltenpoth et al., 2005). The Coriobacterium-specific probe COR653 (5′-fluorescein isothiocyanate labeled) (Harmsen et al., 2000) and the general eubacterial probe EUB338 (5′-Cy3 labeled) (Amann et al., 1990) were used for FISH.

For the whole-mount FISH of the symbionts in the M3 section of the midgut, an adult female P. apterus was dissected, and the M3 was fixated in 70% ethanol. The sample was dehydrated in acetone and then embedded in cold polymerizing resin (Technovit 8100, Heraeus Kulzer) according to the manufacturer's instructions. Sections of 4 μm thickness were made with a diamond knife on a Reichert–Jung 2040 Autocut Microtome. FISH on the embedded tissue sections was carried out without further pretreatment as described above, using the probes COR653 and EUB338 (both 5′-Cy3 labeled), and the Streptomyces-specific probe SPT177 (5′-Cy3 labeled) as a negative control (Kaltenpoth et al., 2006).

Generation of aposymbiotic individuals and reapplication of symbionts

Eggs from 10 different clutches were harvested 4 days after oviposition and randomly divided into two equal parts, one of which served as an untreated control, while the other one was surface sterilized using the procedure described by Prado et al. (2006). Briefly, eggs were treated with 95% ethanol for 5 min, and then submerged for 15–20 s in bleach (12% NaOCl, Aug. Hedinger, Stuttgart, Germany). The residual bleach was washed off with sterile H2O, and the eggs were dried on filter paper. To subsequently avoid C. glomerans contamination, the nymphs that had hatched from both surface-sterilized and control eggs were supplied with autoclaved food (dry linden seeds) and water. The eggs from different clutches and the treated/untreated groups were kept separately and reared to adulthood. The adults were then screened for the presence of C. glomerans using diagnostic PCR.

In a second experiment, 315 eggs from seven different clutches (all oviposited within 24 h) were collected and mixed. Of these, 105 eggs were randomly assigned to an untreated control group. The other 210 were surface sterilized as described above, and 105 of these were randomly assigned to an aposymbiotic control group. The remaining 105 eggs were supplemented with empty eggshells from symbiont-containing individuals to investigate whether symbionts can be taken up from the egg surface. The eggs of each group were reared collectively to adulthood (on autoclaved linden seeds and water) and then screened for the presence of C. glomerans by diagnostic PCR.

To elucidate whether C. glomerans can be acquired from P. apterus feces, fresh feces of adult bugs were collected as follows: dry linden seeds were provided on a Petri dish that had been cleaned with 70% ethanol. On the following day, dried fecal droplets were located on the Petri dish and resuspended with an aliquot of sterile water. For the reapplication experiments of the symbionts, 240 P. apterus eggs from four different clutches were surface sterilized and randomly assigned to six different groups (40 eggs per group). The first group was kept as an aposymbiotic control and reared on autoclaved linden seeds and water, the second group was reared on nonsterile linden seeds and water to investigate whether the symbionts can be taken up from the diet. The eggs in groups 3–5 were treated with P. apterus feces on the second, the fourth, or the sixth day after oviposition, respectively, by applying a suspension of P. apterus feces (in sterile dH2O) onto the egg surface. The eggs of group 6 were kept in the presence of two adult P. apterus females harboring the symbionts. All eggs were reared until adulthood and then screened for the presence of C. glomerans by diagnostic PCR.

Deposition of DNA sequences

Partial 16S rRNA gene sequences for the Coriobacterium symbionts of P. apterus and D. fasciatus were deposited in the GenBank database under the accession numbers FJ554832FJ554837.


Localization of C. glomerans

Coriobacterium glomerans could be reliably detected by diagnostic PCR in the gut samples of both males and females (Table 2). In females, C. glomerans was also detected in most hemolymph samples as well as in 50% of the ovary samples, whereas male hemolymph and testes never gave positive results in Coriobacterium-specific PCRs (Table 2). The detection in the ovary samples, however, may have been due to contaminating cells from the hemolymph. Coriobacterium glomerans was regularly detected both on crushed eggs and in egg washes. FISH experiments showed large numbers of symbionts in the digestive tract (but only in sections M3 and rectum) and in the feces of P. apterus, and lower densities of bacterial cells on the eggs and in egg washes (Fig. 1). The symbionts exhibited the typical chain-like structure as described earlier (Haas & König, 1988). Because of high background fluorescence, no symbiont cells could be detected in the hemolymph samples.

Table 2.   Localization of Coriobacterium glomerans in Pyrrhocoris apterus as revealed by diagnostic PCR (Cor =Coriobacterium)
SamplenCor. positiveCor. positive (%)
 Digestive tract77100
 Digestive tract66100
Figure 1.

 Fluorescence micrograph of Coriobacterium glomerans, the symbiont of Pyrrhocoris apterus. (a–c) Midgut preparations; (d, e) feces; and (f, g) egg washes. Left panel (red): stained with the general eubacterial probe EUB338; right panel (green): stained with the Coriobacterium-specific probe COR653. Scale bars=10 μm.

Whole-mount FISH with the general eubacterial probe EUB338 showed the presence of large numbers of bacteria in the midgut section M3 (Fig. 2a and d), only some of which were identified as C. glomerans by the specific probe COR653 (Fig. 2b and e). Coriobacterium glomerans was mainly located in the anterior part of M3, with cells occurring both attached to the epithelium and free-floating in the gut (Fig. 2b and e). No bacteria were stained by the negative control probe SPT177 (Fig. 2c and f).

Figure 2.

 Whole-mount FISH of Coriobacterium glomerans in the midgut section M3 of Pyrrhocoris apterus with three different fluorescent probes: (a, d) the general eubacterial probe EUB338; (b, e) the Coriobacterium-specific probe COR653; and (c, f) the Streptomyces-specific probe SPT177 (negative control). Scale bars: (a–c), 500 μm; (d–f), 100 μm.

Phylogenetic position of the symbionts

Partial 16S rRNA gene sequences (1340–1380 bp) were obtained for the symbionts of P. apterus and D. fasciatus (three individuals each) and deposited in the GenBank database (accession numbers FJ554832FJ554837). All sequences were most similar to the C. glomerans 16S rRNA gene sequence in the GenBank database. In a phylogeny based on an alignment of 1372 bp of 16S rRNA gene comprising representative actinomycete strains from all different families, the symbiont sequences and C. glomerans formed a well-supported clade within the family Coriobacteriaceae (Fig. 3). Interestingly, the sequence obtained for the P. apterus symbiont differed in about 2.2% of nucleotide positions from the C. glomerans sequence in the GenBank database (accession number X79048). The pairwise distance between the D. fasciatus symbiont and C. glomerans was 3.0%, and that between the sequences of the symbionts of P. apterus and D. fasciatus obtained in this study was about 2.6% (including a 13-bp deletion in the D. fasciatus symbiont). Furthermore, there was variability in the 16S rRNA gene sequences within each symbiont strain: we detected one variable site (A/T) in the partial 16S rRNA gene sequence of the P. apterus symbionts that resulted in double signals at this position (position 446, Streptomyces ambofaciens numbering, see Pernodet et al., 1989) in two of the three individuals. Two more sites were polymorphic (positions 619/620, S. ambofaciens numbering) with two individuals showing two thymines at this position, while the third had two cytosines. Five nucleotide sites were polymorphic in the D. fasciatus symbionts (positions 188, 190, 254, 256, and 257, S. ambofaciens numbering), with two individuals exhibiting one of the two ‘haplotypes,’ respectively, and the third one showing double peaks at all five positions (C/T, A/C, A/G, C/G, and A/G, respectively).

Figure 3.

 Phylogenetic position of Coriobacterium sequences from two pyrrhocorid bugs within the Actinobacteria. Neighbor-joining tree constructed on the basis of 1372 bp of 16S rRNA gene sequences. Bootstrap values (%) were obtained from a search with 100 replicates.

Transmission route of C. glomerans

The presence of C. glomerans on P. apterus eggs and in egg washes strongly suggests vertical transmission of the symbionts via the egg surface. Behavioral observations supported this hypothesis: during oviposition, females regularly touched the eggs with the anus (Fig. 4a and b). Subsequently, dark secretion droplets were visible on the egg surface and accumulated in the places where the eggs were in contact with each other (Fig. 4b and c). Young nymphs were observed probing the egg surface with their proboscis after hatching and thereby probably taking up the symbionts (Fig. 4d).

Figure 4.

 Transmission of symbiotic Coriobacterium glomerans in Pyrrhocoris apterus. The eggs are touched with the anus after oviposition (a, b), and secretion droplets can be seen on the egg surface (b, c). Young nymphs probe the egg surface after hatching (d), presumably for the uptake of symbiotic bacteria.

Surface sterilization of eggs confirmed the symbiont transmission via the egg surface: while in the control group all except one adult P. apterus from 10 different egg clutches harbored C. glomerans (n=124), all individuals were tested negative for the symbionts after surface sterilization of the eggs (n=49; Table 3). The differences in sample sizes in the control and the experimental group were due to higher mortalities in the aposymbiotic group (data not shown).

Table 3.   Effects of surface sterilization of Pyrrhocoris apterus eggs and of the reapplication of symbiont cells via symbiont-containing eggshells or feces on the incidence of Coriobacterium glomerans in adult firebugs
Experiment/groupTreatmentnCor. positiveCor. negative% Cor. positive% Cor. negative
  • *

    Weak PCR bands, probably due to incomplete sterilization of the eggs.

  • Surf., surface; steril., sterilized; symb., symbiotic; apo, aposymbiotic; Cor, Coriobacterium.

Surface sterilization of eggs
 ApoSurf. steril.490490100
Reapplication of symbionts from empty symbiont-containing eggshells
 Control1919 1000
 ApoSurf. steril.7 70100
 Apo+symb.-eggshellsSurf. steril.+symb. eggshells1111 1000
Reapplication of symbionts from symbiont-containing feces
 ApoSurf. steril.5(4)*1(80)*20
 Apo+feces_day2Surf. steril.+feces (day2)1212 1000
 Apo+feces_day4Surf. steril.+feces (day4)99 1000
 Apo+feces_day6Surf. steril.+feces (day6)88 1000
 ApoApo (nonsterile conditions)2 20100
 Apo+adultsSurf. steril.+symb. adults1010 1000

Likewise, in the second experiment, all firebugs in the control group (n=19), but none of the individuals in the aposymbiotic group (n=7), were infected with C. glomerans. Interestingly, all individuals in the third group that had hatched from surface-sterilized eggs but that subsequently had access to empty eggshells from symbiotic bugs were tested positive for the presence of C. glomerans (n=11, Table 3). Thus, symbionts were taken up from the empty eggshells of conspecific individuals.

When surface-sterilized eggs were treated with feces of adult P. apterus, all adult bugs harbored the symbionts, regardless of the time at which the feces were applied to the eggs (2, 4, or 6 days after oviposition; Table 3). Likewise, all individuals hatching from surface-sterilized eggs that were subsequently reared in the presence of symbiont-containing adult P. apterus individuals were tested positive for C. glomerans (Table 3). Aposymbiotic individuals reared on nonsterile linden seeds and water without access to adult bugs or feces did not acquire the symbionts from the environment. However, due to high mortality, the sample size in this group was very small (n=2; Table 3). In the aposymbiotic control group, four specimens showed weak bands in the PCRs, which were probably due to incomplete sterilization of the eggs and the subsequent uptake of some residual symbiont cells from the eggshells by these individuals.


Endosymbiotic bacteria of insects have received considerable attention in the last decades, and many studies have focused on the intimate associations of intracellular symbionts and their hosts and the degree of mutual interdependence of these symbioses. However, recent studies have shown that extracellular gut symbionts of insects can engage in symbiotic interactions of similar intimacy and specificity with their hosts and can experience the same evolutionary and genomic consequences of the symbiotic lifestyle (Hosokawa et al., 2006). In the present study, we elucidated the localization and transmission route of C. glomerans, the specific symbiont associated with the midgut of pyrrhocorid bugs, and we established techniques to successfully create aposymbiotic individuals and to subsequently reapply the symbionts to these aposymbiotic insects.

Cells of C. glomerans were regularly detected in the midgut section M3, the rectum, and in feces of P. apterus. Earlier studies using histological and cultivation-based methods described the bacteria only from the midgut section M3 (Haas & König, 1987). Our results confirm that large numbers of symbiont cells reside in the M3 section, and a portion of these cells passes through the subsequent parts of the digestive tract and is finally excreted. Because the passage through M4 and the ileum occurs very quickly (Silva & Terra, 1994), the symbiont numbers in these portions at any given time may lie below our detection limits in the FISH. In addition to C. glomerans, we detected large numbers of other bacteria in the midgut section M3, which may belong to the genera Streptococcus and Hafnia that have been described earlier from the midgut of P. apterus (Haas & König, 1987).

By detecting the symbionts on the eggs and by generating aposymbiotic individuals via surface sterilization of the eggs, we were able to demonstrate that the symbionts are transmitted vertically via the egg surface. The transfer of symbionts to the eggs probably occurs by smearing of feces onto the egg surface by the female during oviposition. This hypothesis is supported by direct observations of ovipositing females (Fig. 4a–c) and by the results of the reapplication experiments showing that treatment with symbiont-containing feces to the surface of aposymbiotic eggs leads to the acquisition of the symbionts. However, intraovarial application of the bacteria to the surface of the eggs cannot be ruled out at the moment, because the symbionts were also detected in the hemolymph and the ovaries of some females. Haas & König (1987) doubted the transmission route via the egg surface due to the anaerobic nature of the bacteria. On the basis of our observations, we can only speculate on how the symbionts are able to survive outside of the host's gut for the 6–8 days until they are taken up by the hatchlings: (1) the bugs may embed the symbiont cells in a secretion on the egg surface and thereby provide an anaerobic microclimate; or (2) the bacteria differentiate into an aerotolerant and metabolically inactive transfer stage, and growth of the cells is later stimulated in the host's gut.

In addition to the vertical transmission mode of C. glomerans from mother to offspring, our results indicate that the symbionts can also efficiently be acquired horizontally from adult bugs, probably via coprophagy. Because pyrrhocorid bugs often aggregate in large numbers (Socha, 1993), there are ample opportunities for horizontal transfer, and thus, aposymbiotic individuals are probably rare in nature. Evolutionarily, frequent horizontal symbiont transfer may impair host–symbiont coevolution and reduce the effects of the symbiotic lifestyle on mutual adaptations and interdependence (Ferdy & Godelle, 2005). However, some highly specific symbiotic associations are stable even in the absence of vertical transmission, with the symbionts being reliably acquired from the environment in every host generation (McFall-Ngai & Ruby, 1991; Nyholm & McFall-Ngai, 2004; Kikuchi et al., 2007).

Screening for C. glomerans in D. fasciatus and sequencing of the partial 16S rRNA gene showed the presence of closely related symbionts in this species. Thus, C. glomerans may be widespread among pyrrhocorid bugs, and cophylogenetic analyses of several host and symbiont species would be desirable to test for coevolution between the symbiotic partners and to assess the frequency of horizontal transmission events during the evolution of the symbiosis. Interestingly, our results show intraspecific polymorphisms in the symbiont 16S rRNA gene sequence from different host individuals for both species. Potentially, the C. glomerans genome carries multiple copies of the rrn operon that exhibit low levels of sequence heterogeneity (Tourova, 2003). A taxonomic search in the rrnDB (in December 2008) showed 66 actinobacterial genomes containing on average three copies of 16S rRNA genes (http://ribosome.mmg.msu.edu/rrndb/index.php; see Klappenbach et al., 2001; Lee et al., 2009). Alternatively, different symbiont strains may occur in the bug populations and sometimes within the same host individual. However, theoretical considerations suggest that intraindividual competition among symbiont strains may reduce the fitness of the host, and coinfections with multiple strains might therefore be selected against (Frank, 1996, 2003).

Gut symbionts play a vital role in the nutrition of many insects. Although the function of C. glomerans in the midgut of pyrrhocorid bugs is not yet known, it seems likely that the bacteria aid in the degradation of the diet or supply essential nutrients to the host that the insect can neither synthesize itself nor obtain from the diet in sufficient quantities (Haas & König, 1987). Symbiotic supply of essential amino acids and/or vitamins to the host has been reported for the gammaproteobacterial symbionts of several insect taxa, for example, aphids (e.g. Douglas, 1998, 2006; Gündüz & Douglas, in press), tse-tse flies (Nogge, 1981; Akman et al., 2002) and carpenter ants (Gil et al., 2003; Feldhaar et al., 2007), and notably, for the actinobacterial gut symbionts of the triatomine bug Rhodnius prolixus (Baines, 1956; Harington, 1960; Lake & Friend, 1968). Elimination of symbionts in P. apterus leads to retarded growth and a significant increase in mortality (S. Winter & M. Kaltenpoth, unpublished data), which is consistent with an essential function of the symbionts for the nutrition of the host. However, further studies are needed to elucidate the metabolic activity of C. glomerans and its importance for the nutrition of pyrrhocorid bugs.


We are grateful to Dieter Mahsberg and Annett Endler for kindly providing D. fasciatus individuals and for help with the identification of the species. We thank Margot Schilling for the embedding and sectioning of firebug midguts for the whole-mount FISH. We gratefully acknowledge funding by the Volkswagen foundation (VW) through a postdoctoral fellowship for M.K.