1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Symbiotic bacteria of the genus Verminephrobacter (Betaproteobacteria) were detected in the nephridia of 19 out of 23 investigated earthworm species (Oligochaeta: Lumbricidae) by 16S rRNA gene sequence analysis and fluorescence in situ hybridization (FISH). While all four Lumbricus species and three out of five Aporrectodea species were densely colonized by a mono-species culture of Verminephrobacter, other earthworm species contained mixed bacterial populations with varying proportions of Verminephrobacter; four species did not contain Verminephrobacter at all. The Verminephrobacter symbionts could be grouped into earthworm species-specific sequence clusters based on their 16S rRNA and RNA polymerase subunit B (rpoB) genes. Closely related host species harboured more closely related symbionts than did distantly related hosts. Co-diversification of the symbiotic partners could not be demonstrated unambiguously due to the poor resolution of the host phylogeny [based on histone H3 and cytochrome c oxidase subunit I (COI) gene sequence analyses]. However, there was a pattern of symbiont diversification within four groups of closely related hosts. The mean rate of symbiont 16S rRNA gene evolution was determined using a relaxed clock model, and the rate was calibrated with paleogeographical estimates of the time of origin of Lumbricid earthworms. The calibrated rates of symbiont 16S rRNA gene evolution are 0.012–0.026 substitutions per site per 50 million years and thus similar to rates reported from other symbiotic bacteria.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Earthworms (Oligochaeta: Lumbricidae) harbour symbiotic bacteria in their excretory organs, the nephridia (Knop, 1926; Schramm et al., 2003). The nephridia are found in pairs in each segment of the worm and consist of a long coiled tube leading from the coelomic cavity through three major loops finally exiting the body wall via an exterior pore (Fig. 1A). The passing of fluids from the coelom to the exterior plays an important role in both osmoregulation and excretion of nitrogenous waste (reviewed in Edwards and Bohlen, 1996). In the six earthworm species investigated so far the symbiotic bacteria are confined to a specific part of the second loop called the ampulla, where they form a dense population lining the lumen wall. Furthermore, the symbionts were species-specific and formed a monophyletic group within the genus Acidovorax (Betaproteobacteria) (Schramm et al., 2003). Recently, the genus Verminephrobacter was created to accommodate this assembly of symbionts (Pinel et al., 2008). For the worm Eisenia fetida, it was shown that the symbionts are transmitted vertically through the cocoon where they are deposited along with eggs and sperm (Davidson and Stahl, 2006). The fertilization takes place in the cocoon, and as the embryo develops the symbionts migrate into the ampulla part of the nephridia. Colonization takes place during a critical time period in early development and after this window of opportunity the symbionts can no longer colonize the nephridia (Davidson and Stahl, 2006; Davidson and Stahl, 2008). Symbionts can be isolated in pure culture (Pinel et al., 2008; M.B. Lund, S. Schätzle, K.U. Kjeldsen and A. Schramm, submitted), and symbiont-free worms can grow and reproduce in laboratory culture (Davidson and Stahl, 2006), thus the symbiotic partners are not strictly dependent on each other.


Figure 1. A. Schematic outline of nephridia in an earthworm. Middle diagram: dissected earthworm with a pair of nephridia attached to the body wall in each segment. Right diagram: detail of single nephridium showing the three major loops. The symbionts are restricted to the ampulla (red). B–F. FISH images of nephridial bacteria in the ampulla of five different earthworm species after double hybridization with the eubacterial probe EUB338-CY5 (red, other bacteria) and the Verminephrobacter-specific probe LSB145-CY3 (green, Verminephrobacter appear yellow). Green is tissue autofluorescence (only B–D). Lumbricus terrestris (B) and Aporrectodea tuberculata (C) are exclusively inhabited by Verminephrobacter. Aporrectodea rosea (D), Dendrobaena veneta (E) and Allolobophora chlorotica (F) have a mixed bacterial population in their nephridia. (C) shows how the symbionts attach to the wall of the ampulla. (B) and (C) show the abrupt border, between the ampulla and the middle tubule, that the Verminephrobacter never crosses based on microscopic observations. The white arrows indicate direction of urine flow. Scale bar is 20 μm.

Download figure to PowerPoint

Based on the known monophyly, species specificity and vertical transmission of the symbionts, we hypothesized that earthworms and their nephridial symbionts have co-diversified.

The objectives of this study were (i) to test for the ubiquity and species specificity of Verminephrobacter in a larger collection of lumbricid earthworm species, (ii) to evaluate possible co-diversification of Verminephrobacter with their earthworm hosts, and (iii) because the time of origin of lumbricid earthworm can be inferred from paleogeographical data, to assess the evolutionary rate of symbiont 16S rRNA genes.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Symbiotic bacterial populations in lumbricid earthworms

Adult individuals of 23 lumbricid earthworm species, collected in Europe and the USA (Table 1), were analysed for the presence of Verminephrobacter nephridial symbionts by fluorescence in situ hybridization (FISH) and semi-specific PCR. Although Verminephrobacter spp. were detected in almost all worms, many earthworm species also contained other bacteria in their ampullas (Fig. 1). The Verminephrobacter symbionts formed a dense layer of attached cells on the wall of the ampulla lumen (Fig. 1C), whereas the other bacteria often appeared more loosely attached. Based on these observations, earthworms can be subdivided into three categories: (i) species that were exclusively populated by Verminephrobacter, i.e. the genus Lumbricus (Fig. 1B), and a monophyletic group of Aporrectodea species [Atuberculata (Fig. 1C), A. caliginosa and A. longa], (ii) species with a mixed bacterial population with varying relative abundance of Verminephrobacter, e.g. in Aporrectodea rosea (Fig. 1D) and Dendrobaena veneta (Fig. 1E), they accounted for ≥ 50% of the nephridial inhabitants, whereas in Allolobophora chlorotica the Verminephrobacter symbionts were vastly outnumbered (Fig. 1F), and (iii) species without Verminephrobacter but with other bacteria present (Fig. 2), i.e. four of the five investigated species of the polyphyletic genus Dendrobaena; absence of Verminephrobacter was confirmed in 3–16 individuals per species by semi-specific PCR and/or cloning and FISH. Among the species that do harbour Verminephrobacter symbionts, individuals without Verminephrobacter have never been found among the more than 50 worms sampled (i.e. two to six individuals per species).

Table 1.  Species list, sampling location and genes sequenced.
  • a.

    ns = no symbiont; absence confirmed in 3–16 individuals, for details see Experimental procedures.

  • Accession numbers are given for all sequences in the analysis.

Allolobophora chlorotica (Savigny)Ach3Svendborg, DenmarkFJ214210FJ374780FJ214195
Ach5Svendborg, DenmarkFJ214231FJ214261FJ214204
Aporrectodea caliginosa (Savigny)Ac9 culture FJ214202
Acc2Svendborg, DenmarkFJ214218FJ214248FJ214188FJ214281, FJ214282
Acc3Svendborg, DenmarkFJ374784FJ214189
Aporrectodea icterica (Savigny)Ai1Aarhus, DenmarkFJ214220FJ214249FJ214197
Ai3Aarhus, DenmarkFJ214219FJ374781FJ214196
Aporrectodea longa (Ude)Al3Aarhus, DenmarkFJ214222FJ214251FJ214190FJ214284, FJ214285
Al4Aarhus, DenmarkFJ214221FJ214250FJ214191FJ214286, FJ214287
Al7Aarhus, DenmarkFJ892716, FJ892717
Aporrectodea rosea (Savigny)Ar1Aarhus, DenmarkFJ374777FJ214252FJ214194FJ214288, FJ214302
Ar5Svendborg, DenmarkFJ214232FJ214262FJ214203FJ214283, FJ214308
Aporrectodea tuberculata (Eisen)At10Konnevesi, FinlandFJ214268FJ214187
At2Aarhus, DenmarkFJ214223FJ374783FJ374774FJ214289, FJ214290
At4 culture FJ214186FJ214311
4.2-9Seattle, Washington, USAAJ543437
4.3-1Seattle, Washington, USAAJ543438
Dendrobaena attemsi (Michaelsen)DaOlympic National Forest, Washington, USAFJ214224FJ214254nsans
Dendrobaena clujensisDcCluj, RomaniaFJ374778FJ214253nsns
Dendrobaena hortensisEzVermiculture, Florida, USAFJ214229FJ214259nsns
Dendrobaena octaedra (Savigny)D16Disco, GreenlandFJ214234nsns
DMUSilkeborg, DenmarkFJ214235nsns
Dendrobaena veneta (Bouché)Dv1Vermiculture, the NetherlandsFJ214233FJ214198FJ214292
Dv2Vermiculture, the NetherlandsFJ214238FJ214199FJ214293, FJ214294
Dv3Vermiculture, the NetherlandsFJ214264
Dendrodrilus rubidus (Savigny)Dr1Aarhus, DenmarkFJ214182FJ214295, FJ214296
Dr10Konnevesi, FinlandFJ214209FJ214239FJ214183
Dr2Aarhus, DenmarkFJ374776FJ374782FJ214184FJ214297, FJ214309
Eisenia fetida (Savigny)EfSeattle, Washington, USAFJ214228FJ214258DC093612
Ef1Svendborg, DenmarkFJ214217FJ214247FJ214179FJ214298, FJ214299
Ef2Svendborg, DenmarkFJ214216FJ214246FJ214180FJ214273, FJ214300
Ve genome CP000542CP000542
EfNe1Seattle, Washington, USAAJ543439
Eisenia lucensElAlba, RomaniaFJ214225FJ214255FJ214181
Eisenoides carolinensisEcStateline, Louisiana, USAFJ214226FJ214256FJ214185
Helodrilus oculatus (Hoffmeister)Ho2Aarhus, DenmarkFJ374775FJ214245FJ214192FJ214303
Ho3Aarhus, DenmarkFJ214271FJ214193FJ214304, FJ214310
Lumbricus castaneus (Savigny)Lc1Svendborg, DenmarkFJ214215FJ214244FJ214175FJ214274, FJ214275
Lc2Svendborg, DenmarkFJ214237FJ214272FJ214176FJ214276, FJ214301
Lc21Aarhus, DenmarkFJ892723
Lc22Aarhus, DenmarkFJ892722
Lumbricus festivus (Savigny)Lf1Aarhus, DenmarkFJ214214FJ214243FJ214170
Lf2Svendborg, DenmarkFJ214174FJ892720, FJ892721
Lf3Svendborg, DenmarkFJ214270FJ214173FJ214277, FJ214307
Lumbricus rubellus (Savigny)LRMorgantown, Kentucky, USAFJ214227FJ214257FJ214178
Lr1Silkeborg, DenmarkFJ214213FJ214242FJ214177FJ214279, FJ214280
Lr12Konnevesi, FinlandFJ892718, FJ892719
Lumbricus terrestris (Linnaeus)LTBremerton, Washington, USAFJ214230FJ214260FJ214171
11.2-15Bayreuth, GermanyAJ543435
4.7-1Seattle, Washington, USAAJ543436
Lt1Aarhus, DenmarkFJ214212FJ214241
Lt2Aarhus, DenmarkFJ214211FJ214240FJ214172
Lt6Aarhus, DenmarkFJ214278, FJ214291
Octolasion cyaneum (Savigny)Oc1Cineba, Washington, USAFJ214208
Octolasion lacteumOl1Szentmargitfalva, HungaryFJ214236FJ214267FJ214207
Octolasion tyrtaeum (Savigny)Osp.Olympic National Forest, Washigton, USAFJ374779FJ214263FJ214201
Ot1Aarhus, DenmarkFJ214269FJ214200FJ214305, FJ214306
Ot3Aarhus, DenmarkFJ214265FJ214205
Ot6Bayreuth, GermanyFJ214266FJ214206

Figure 2. A. Comparative phylogenies of earthworms and their nephridial symbionts. Left: symbiont phylogeny based on 16S rRNA genes. Right: earthworm phylogeny based on concatenated H3 and COI nucleotide sequences. B. Symbiont phylogeny based on rpoB nucleotide sequences. All trees are strict consensus trees from NJ, ML and BI. All branch lengths are based on BI. Numbers on branches are posterior probabilities given by MrBayes. Numbers inside groups indicate the number of earthworm individuals contained in the group. Species that do not harbour Verminephrobacter are marked by an asterisk. Species in bold are exclusively inhabited by Verminephrobacter in their nephridia. Full trees can be seen in Figs S1–S3.

Download figure to PowerPoint

Host specificity of Verminephrobacter symbionts

Out of the 23 examined earthworm species 19 had Verminephrobacter symbionts in their nephridia. For each earthworm species two to six individuals were examined and for eight of the species (A. rosea, A. tuberculata, Dendrodrilus rubidus, E. fetida, Lumbricus festivus, Lumbricus rubellus, Lumbricus terrestris and Octolasion tyrtaeum) individuals were collected from different geographical locations (Table 1). The bacterial clones from different individuals of the same earthworm species formed their own cluster in both the 16S rRNA gene- and RNA polymerase subunit B (rpoB) gene-based phylogenies (Fig. 2), except for the symbionts of the closely related Lumbricus spp.; these symbionts could not be separated phylogenetically based on 16S rRNA gene sequences but only when using the more informative rpoB gene. The grouping of Verminephrobacter nephridial symbionts obtained from the same earthworm species indicates that the symbiosis is species-specific. Whether the species specificity is solely a consequence of vertical symbiont transfer (Davidson and Stahl, 2006), or governed by a host-symbiont recognition mechanism, is unknown at present.

Comparative phylogeny of earthworm hosts and Verminephrobacter symbionts

To test for co-diversification patterns in the evolution of earthworms and their nephridial symbionts, the phylogeny of the hosts, based on histone H3 (H3) and cytochrome c oxidase subunit I (COI) gene sequences, was compared with that of the symbionts based on 16S rRNA and rpoB gene sequences (Fig. 2). Using concatenated H3 and COI nucleotide sequences, the phylogenetic relationship of the earthworms could only be resolved among closely related species. The deeper branching pattern of the tree remained unresolved and has thus been drawn as a multifurcation with smaller resolved clades (Fig. 2A). Lumbricid earthworm classification has until recent years been based on morphological data only, and this classification is not supported by the molecular phylogeny in this study; the only morphologically defined genera confirmed by the molecular approach were Lumbricus and Octolasion, whereas the genus Dendrobaena remained unresolved, and Aporrectodea and Eisenia were rendered polyphyletic. A similar discrepancy between molecular and morphological classification has previously been found in a molecular phylogeny study of earthworms based on 16S and 18S rRNA, and COI gene sequences (Pop et al., 2007).

For the symbiont, the phylogenies based on 16S rRNA (Fig. 2A) and rpoB (Fig. 2B) gene sequences had only minor incongruencies and rpoB gave a higher resolution of closely related species than 16S rRNA, as found for most bacteria (Case et al., 2007). For example, 16S rRNA gene sequence analysis failed to resolve both the genus Lumbricus and the group of A. caliginosa, A. tuberculata, A. longa, A. rosea and Helodrilus oculatus (Fig. S1).

The degree of congruency of the earthworm host and symbiont phylogenies could not be assessed in detail due to the low resolution of particularly the earthworm phylogeny. The only four groups that could be resolved in the earthworm tree were however also found in the symbiont 16S rRNA tree. These groups were: (i) the genus Lumbricus, (ii) the genus Octolasion, (iii) the clade comprised of A. caliginosa/A. tuberculata/A. longa, and (iv) the clade comprised of Aporrectodea icterica/Allolobophora chlorotica . When comparing the earthworm phylogeny with the more finely resolved rpoB-based symbiont phylogeny, the same overall pattern was seen; within the group of A. caliginosa/A. tuberculata/A. longa the tree topologies were identical. Although the symbionts of the genus Lumbricus formed a monophyletic group, their relationships inferred from rpoB were not fully congruent with that of the earthworm hosts. The groups resolved in the host and symbiont phylogenies indicate that the symbionts have diversified within groups of closely related hosts on at least four separate occasions. Whether true co-diversification of hosts and symbionts has occurred during the evolution of lumbricid earthworms is not evident from the current data but will require a better resolved, multigene, host phylogeny.

An alternative to co-diversification would be frequent horizontal transfer followed by speciation of the symbionts; in that case, worm species co-occurring in nature were expected to carry more closely related symbionts than species with differing habitats or distribution. However, there was no overall pattern in the preferred habitat of the host and the similarity of the Verminephrobacter symbionts they carry; for example, the earthworms E. fetida and D. veneta are both compost worms commonly bred in vermiculture, but their Verminephrobacter symbionts are very different. Along the same line, the earthworms L. terrestris and A. longa are both anecic worms living in vertical burrows in the soil but carry very different symbionts. In addition, the Verminephrobacter symbionts are not able to survive in the soil environment; Verminephrobacter have never been detected in soil samples nor in bedding from laboratory worm cultures (Pinel et al., 2008) and earthworms are therefore not expected to encounter symbionts of other earthworms in high numbers. Thus, it is unlikely that the symbionts undergo frequent horizontal transfers.

None of the Dendrobaena species, except D. veneta, harbour Verminephrobacter. Dendrobaena species form two separate, unresolved clades in the earthworm tree (Fig. 2A). Either the loss of symbionts must have happened once in the genus Dendrobaena, followed by a later horizontal acquisition by D. veneta, or the symbionts were lost in multiple events. The earthworms without Verminephrobacter symbionts all harbour other FISH-detectable bacteria in their nephridia, which may have replaced the Verminephrobacter.

Rates of molecular evolution in symbiont 16S rRNA

The almost universal occurrence of Verminephrobacter symbionts in Lumbricidae together with their absence in other earthworm families (S.K. Davidson and S. James, unpubl. data), and the monophyly of the symbionts, suggests that the Verminephrobacter–earthworm association originated in the most recent common ancestor of lumbricid earthworms. Earthworms consist mainly of soft tissues, and have therefore not left any fossil record; however, the time of origin of lumbricid earthworms can be inferred from paleogeographical information (Bouché, 1983); this offers a rare opportunity to calibrate bacterial evolutionary rates. Based on the original (i.e. pre-human-impacted) distribution of lumbricid earthworms in Laurasia only, the family Lumbricidae has been estimated to originate during the split of Laurasia in the Cretaceous, i.e. 62–136 million years ago (Bouché, 1983; Roest and Srivastava, 1989; Pe-Piper et al., 2007). It appears likely that Verminephrobacter have been associated with lumbricid earthworms ever since.

A relaxed molecular clock model was used to calculate the mean substitution rate in 16S rRNA in the symbiont clade. The mean rate was 0.0319 substitutions per site with a 95% highest posterior density interval of (0.0235; 0.0402). Calibrating the mean rate with the estimated origin of lumbricid earthworms gave an evolutionary rate of 0.0117 (0.0087; 0.0148) to 0.0257 (0.0190; 0.0324) substitutions per site per 50 million years (numbers in brackets are calculated rates for the 95% credibility set). These estimated evolutionary rates are in the range reported from endosymbionts of woodroaches (0.0084–0.0111 substitutions per site per 50 million years; Maekawa et al., 2005) and aphids (0.0076–0.0232 substitutions per site per 50 million years; Moran et al., 1993). Endosymbiotic bacteria in aphids exhibit accelerated rates of 16S rRNA gene evolution relative to closely related free-living bacteria (Moran, 1996). Accelerated rates have also been found in maternally transmitted sulfur-oxidizing endosymbionts of solemyid bivalves (Peek et al., 1998). Substitution rates are higher in populations with a small effective population size (Ne) according to the nearly neutral theory of molecular evolution (Ohta, 1992). Thus the accelerated substitution rates found in endosymbionts are attributed to the low Ne that arises as a consequence of recurring bottlenecks in symbiont transmission between host generations (Moran, 1996).

We do not know whether the 16S rRNA substitution rates in Verminephrobacter are in fact accelerated compared with closely related, free-living Acidovorax species. However, since the evolutionary rate in earthworm symbionts compares to that of other symbionts with accelerated rates they may experience the same severe bottlenecks in transmission from one host generation to the other. At present, the number of symbionts that are transferred to the next generation, i.e. deposited in the cocoon albumin and finally infecting the embryo (Davidson and Stahl, 2006; 2008), is unknown. Our results suggest, however, that the founding population must be rather small.

In conclusion, Verminephrobacter is present and host species-specific in 19 out of the 23 lumbricid earthworms examined in this study. There is a pattern of symbiont diversification within groups of closely related hosts; however, it remains unresolved whether the symbiotic partners have truly co-diversified. The rates of 16S rRNA evolution compare to those of other symbionts, indicating that the symbionts experience a bottleneck during transmission from one host generation to the next.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Collection and preparation of samples

Adult individuals of 23 different earthworm species were collected in Europe, Greenland and the USA in 2005–2008 (Table 1) and determined to species by analytical keys (Sims and Gerard, 1985; Andersen, 1997). The worms were cleaned with water and killed in 70% ethanol. Nephridia from 20–30 segments were dissected out using a dissection microscope and immediately frozen at −20°C and stored for later DNA extraction. For FISH, the worms had their gut removed and were fixed in 4% (w/v) para-formaldehyde for 2–3 h, washed twice in phosphate-buffered saline and stored in 70% ethanol at −20°C.

Fluorescence in situ hybridization (FISH)

The presence of Verminephrobacter spp. and other bacteria in the nephridia was detected by FISH in two to three individuals of each species. Hybridizations were carried out directly on fixed, ethanol-stored earthworm body wall with the nephridia still attached. The probes: EUB338-MIX-CY5 (Daims et al., 1999), targeting all bacteria, and LSB145-CY3 (Schweitzer et al., 2001), targeting some Acidovorax spp. and all Verminephrobacter spp., were used in combination. Hybridizations were carried out at 35% formamide for 2–4 h according to published protocols (Pernthaler et al., 2001). After hybridization and washing, the tissue fragments were counter-stained with 5 μg μl−1 DAPI (4′,6-diamidino-2-phenylindole) for 30 min on ice and rinsed twice in water. For large earthworm individuals single nephridia were dissected from the body wall and mounted on a glass slide prior to microscopy. Smaller earthworms were mounted directly on glass slides. Samples were analysed with an Axiovert 200M epifluorescence microscope equipped with an ApoTome device for optical sectioning (Carl Zeiss, Jena, Germany).

DNA extraction, PCR and sequencing

DNA was extracted from nephridia using the DNeasy Blood and Tissue kit (Qiagen). For DNA extraction and gene amplification by PCR, two to five individuals from each species were used. DNA extracts were used for PCR amplification of two earthworm genes: COI and Histone H3, and two symbiont genes: 16S rRNA and rpoB. All PCR reactions were run with Taq master mix (Ampliqon) containing 1.5 mM MgCl2, 0.2 mM dNTP, 0.025 U μl−1 Ampliqon Taq DNA polymerase and 0.2 μM of each primer. Earthworm mitochondrial COI was PCR amplified using the primer pair LCO1490 and HCO2192 (Folmer et al., 1994) giving a fragment of 710 bp. The PCR was run with the following thermal cycling: 94°C for 5 min, 35 cycles of 30 s at 94°C, 30 s at 49°C, 1 min at 72°C, final elongation for 5 min at 72°C. The PCR product was cloned directly using pGEM®-T easy vector system from Promega and transformed into competent Eschericha coli JM109. M13 forward and reverse vector primers were used for sequencing.

Earthworm nuclear Histone H3 were PCR amplified using the primers H3F and H3R (Colgan et al., 2000) resulting in a fragment of 399 bp. Touchdown PCR was run with the following thermal cycling: 93°C for 5 min followed by 30 cycles of 93°C for 30 s, 55°C for 1 min and 72°C for 1 min, the annealing temperature was lowered by 1°C during the first seven cycles ending with an annealing temperature of 48°C in the last 22 cycles; final elongation was at 72°C for 5 min. PCR products were purified (GenEluteTM PCR Clean-Up kit, Sigma) and sequenced directly using H3F and H3R as sequencing primers.

Bacterial 16S rRNA genes were PCR amplified with the universal bacterial primers 26F and 1492R (Lane, 1991). PCR was run with the following cycles: 93°C for 5 min followed by 25 cycles of 93°C for 30 s, 57°C for 1 min and 72°C for 1:30 min. Cycling ended with 72°C for 5 min. The PCR products were cloned directly using either pGEM®-T easy vector system (Promega) or TOPO® XL PCR Cloning kit (Invitrogen). Clones were screened for the right insert with a semi-specific colony PCR targeting the nephridial symbionts using the primers LSB145 (Schweitzer et al., 2001) and 1492R. The PCR was run with the same thermal cycling as for 16S rRNA gene amplification except the annealing temperature was lowered to 50°C. M13 forward and reverse vector primers were used for sequencing.

Bacterial rpoB was PCR amplified with the primers rpoB43F (5′-ttcggcasccgcgacagcgygc) and rpoB4061R (5′-cagkgarcggatttccttgacca) resulting in a 4020 bp fragment. The primers were designed based on full-length rpoB sequences retrieved from fully sequenced genomes ( to specifically target the rpoB genes of Verminephrobacter and closely related Acidovorax species. The thermal cycling was run with 1 min at 94°C followed by 25 cycles of 30 s at 94°C, 1 min at 60°C and 5 min at 72°C, final elongation 10 min at 72°C. The right-sized band was gel extracted and cloned. For some samples the gel extract was re-amplified to yield enough product for cloning with the pGEM®-T easy vector system (Promega). M13 forward and reverse vector primers and one internal forward primer were used for sequencing resulting in good sequence data of the first 1463 bp and the last 703 bp of each sequence, 2166 nucleotides in total.

All sequencing was carried out at Macrogen, Korea; GenBank accession numbers of the sequences retrieved in this study are listed in Table 1.

Earthworm species that were found not to contain Verminephrobacter were double-checked as follows: Dendrobaena octaedra was analysed by applying the semi-specific PCR screening procedure, also used to screen the clone libraries, directly on nephridial DNA extracts. The semi-specific PCR was applied to 13 individuals collected at 11 different sites in: Canada (Lethbridge and Dunvegen), Finland (Jyväskyla), Poland (Glogow and Olkusz), Sweden (Lund, Umeå and Uppsala), Greenland (Disco and Nassasuaq) and Denmark (Silkeborg). Individuals from Denmark and Greenland were also analysed by FISH. For the earthworm species Dendrobaena attemsi, Dendrobaena clujensis and Dendrobaena hortensis three individuals of each species were analysed in three ways: (i) sequencing of 96 clones, (ii) semi-specific PCR screening procedure used directly on the nephridial DNA extracts, and (iii) FISH conducted with the specific probe LSB145. All tests confirmed the absence of Verminephrobacter in the four Dendrobaena species.

Phylogenetic analysis

The earthworm phylogeny was based on concatenated H3 and COI nucleotide sequences. The sequences were unambiguously aligned using clustalx (Thompson et al., 1997) and the software package ARB (Ludwig et al., 2004). The number of parsimony informative sites in COI and H3 were 261 out of 658 and 105 out of 328 respectively. The tree (Fig. 1A) is a strict consensus tree based on ARB neighbour joining (NJ) and ARB AxML maximum likelihood (ML) and Bayesian inference (BI) using MrBayes (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003); the numbers on the branches are the posterior probabilities from BI. For a few of the species, only either COI data or H3 data were available. When using the NJ algorithm, these sequences were first excluded from the analysis and then added to the NJ tree using the ‘ARB parsimony Quick add marked’ function in combination with appropriate filters correcting for the shorter sequence length. For BI, the data were partitioned in six parts consisting of the three codon positions in both H3 and in COI. The HKY model with equal rates of substitution along the sequence length was applied. Rates of substitution and transition/transversion ratios were allowed to vary across partitions. Two independent MCMC analyses were run in 100 000 generations each. Convergence of the runs was verified by plotting estimated parameters against number of generations.

The bacterial phylogeny was based on 16S rRNA gene sequences and supplemented with rpoB nucleotide sequences from most of the groups to improve the resolution of the phylogeny. The 16S rRNA gene sequences were aligned against an existing ARB database (ssu_jan04). The number of parsimony informative sites in 16S rRNA gene sequences was 130 out of 1466 in total. The strict consensus tree (Fig. 2A) was based on NJ and ML calculated in ARB and BI using MrBayes. For BI, the HKY model of nucleotide substitution and a gamma-distributed rate heterogeneity model with a proportion of invariant sites (HKY + I + Γ) were used. The posterior probabilities are displayed on the tree branches. Two independent MCMC analyses were run in 1 000 000 generations and convergence of runs was verified by plotting all estimated parameters against number of generations. The rpoB-based tree (Fig. 2B) was constructed in the same way as the 16S rRNA-based tree except that the model used in BI was HKY + _ + Γ. The number of parsimony informative sites in rpoB gene sequences was 619 out of 2167 in total.

Rates of molecular evolution in symbiont 16S rRNA

The very uneven branch lengths in the symbiont 16S rRNA tree (Fig. 1A) alone indicate that there is no rate constancy in the symbiont cluster and hence a strict molecular clock model will not fit the data. Therefore, to estimate the rate of molecular evolution in 16S rRNA for the Verminephrobacter clade, a relaxed clock model was implemented using the program beast v1.4.8 (Drummond and Rambaut, 2007). Only the 16S rRNA sequences from the symbionts were included in the analysis. An HKY model of nucleotide substitution with gamma-distributed among-site rate heterogeneity with a proportion of invariant sites (HKY + I + Γ), was used along with an uncorrelated log-normal distribution (UCLN) model of rate variation among branches (relaxed clock) (Drummond et al., 2006). In addition, a Yule prior on speciation rates was used. The total height of the tree was fixed to 1 by setting the prior on the treeModel.rootHeight to a normal distribution with a mean of 1 and a standard deviation of 0.0001. Three independent MCMC analyses were run in 10 000 000 generations each. The program Tracer v1.4 (Rambaut and Drummond, 2007) was used to determine convergence and proper mixing of the MCMC runs. Tracer v1.4 was also used to combine the three independent MCMC runs to obtain estimates of the mean value, along with the 95% credible set of mean substitution rates in the symbiont cluster. Since the bacterial 16S rRNA substitution rate estimated in beast was calibrated to 1, the rate could be calibrated with lumbricid earthworm emergence time by dividing the mean substitution rate estimate with the proposed time of origin of lumbricid earthworms, i.e. 62–136 million years ago (Bouché, 1983; Roest and Srivastava, 1989; Pe-Piper et al., 2007).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank Britta Poulsen and Ryan Powell for expert help in the laboratory, and Finn Borchsenius for help with calculating evolutionary rates. Markus Horn is acknowledged for providing the worm O. tyrtaeum from Germany. This study was financially supported by the Danish Research Council for Natural Sciences (Grant 21-04-0410 to A.S.) and US National Science Foundation DEB-0516520 to S.K.D. and D.A.S. and DEB-0516439 to S.J.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information
  • Andersen, C. (1997) Regnorme. Natur og Museum 36: 4. Naturhistorisk Museum, Aarhus, Denmark.
  • Bouché, M.B. (1983) The establishment of earthworm communities. In Earthworm Ecology – From Darwin to Vermiculture. Satchell, J.E. (ed.). New York, USA: Chapman and Hall, pp. 431448.
  • Case, R.J., Boucher, Y., Dahllöf, I., Holmström, C., Doolittle, W.F., and Kjelleberg, S. (2007) Use of 16S rRNA and rpoB genes as molecular markers for microbial ecology studies. Appl Environ Microbiol 73: 278288.
  • Colgan, D.J., Ponder, W.F., and Eggler, P.E. (2000) Gastropod evolutionary rates and phylogenetic relationships assessed using partial 28S rDNA and histone H3 sequences. Zool Scripta 29: 2963.
  • Daims, H., Bruhl, A., Amann, R., Schleifer, K.H., and Wagner, M. (1999) The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: development and evaluation of a more comprehensive probe set. Syst Appl Microbiol 22: 434444.
  • Davidson, S.K., and Stahl, D.A. (2006) Transmission of nephridial bacteria of the earthworm Eisenia fetida. Appl Environ Microbiol 72: 769775.
  • Davidson, S.K., and Stahl, D.A. (2008) Selective recruitment of bacteria during embryogenesis of an earthworm. ISME J 2: 510518.
  • Drummond, A.J., and Rambaut, A. (2007) beast: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 214.
  • Drummond, A.J., Ho, S.Y., Phillips, M.J., and Rambaut, A. (2006) Relaxed phylogenetics and dating with confidence. PLoS Biol 4: e88.
  • Edwards, C.A., and Bohlen, P.J. (1996) Biology and Ecology of Earthworms. London, UK: Chapman & Hall.
  • Folmer, O., Black, M., Hoeh, W., Lutz, R., and Vrijenhoek, R. (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol 3: 294299.
  • Huelsenbeck, J.P., and Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754755.
  • Knop, J. (1926) Bakterien und Bakteroiden bei Oligochäten. Z Morph Ökol Tiere 6: 588624.
  • Lane, D.J. (1991) 16S/23S rRNA sequencing. In Nucleic Acid Techniques in Bacterial Systematics. Stackebrandt, E., and Goodfellow, M. (eds). New York, NY, USA: John Wiley and Sons, pp. 115175.
  • Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, et al. (2004) ARB: a software environment for sequence data. Nucleic Acids Res 32: 13631371.
  • Maekawa, K., Park, Y.C., and Lo, N. (2005) Phylogeny of endosymbiont bacteria harbored by the woodroach Cryptocercus spp. (Cryptocercidae: Blattaria): molecular clock evidence for a late Cretaceous–early Tertiary split of Asian and American lineages. Mol Phylogenet Evol 36: 728733.
  • Moran, N.A. (1996) Accelerated evolution and Muller's rachet in endosymbiotic bacteria. Proc Nat Acad Sci USA 93: 28732878.
  • Moran, N.A., Munson, M.A., Baumann, P., and Ishikawa, H. (1993) A molecular clock in endosymbiotic bacteria is calibrated using the insect hosts. Proc R Soc Lond B 253: 167171.
  • Ohta, T. (1992) The nearly neutral theory of molecular evolution. Annu Rev Ecol Syst 23: 263286.
  • Peek, A.S., Vrijenhoek, R.C., and Gaut, B.S. (1998) Accelerated evolutionary rate in sulfur-oxidizing endosymbiotic bacteria associated with the mode of symbiont transmission. Mol Biol Evol 15: 15141523.
  • Pe-Piper, G., Piper, D.J.W., Jansa, L.F., and De Jonge, A. (2007) Early cretaceous opening of the north Atlantic ocean: implications of the petrology and tectonic setting of the Fogo seamounts off the SW grand banks, Newfoundland. Geol Soc Am Bull 119: 712724.
  • Pernthaler, J., Glöckner, F.O., Schonhuber, W., and Amann, R. (2001) Fluorescence in situ hybridization (FISH) with rRNA-targeted oligonucleotide probes. Methods Microbiol 30: 207226.
  • Pinel, N., Davidson, S.K., and Stahl, D.A. (2008) Verminephrobacter eiseniae gen. nov., sp. nov., a nephridial symbiont of the earthworm Eisenia foetida (Savigny). Int J Syst Evol Microbiol 58: 21472157.
  • Pop, A.A., Cech, G., Wink, M., Csuzdi, C., and Pop, V.V. (2007) Application of 16S, 18S rDNA and COI sequences in the molecular systematics of the earthworm family Lumbricidae (Annelida, Oligochaeta). Eur J Soil Biol 43: S43S52.
  • Rambaut, A., and Drummond, A.J. (2007) Tracer v1.4[WWW document]. URL
  • Roest, W.R., and Srivastava, S.P. (1989) Sea-floor spreading in the Labrador Sea – a new reconstruction. Geology 17: 10001003.
  • Ronquist, F., and Huelsenbeck, J.P. (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 15721574.
  • Schramm, A., Davidson, S.K., Dodsworth, J.A., Drake, H.L., Stahl, D.A., and Dubilier, N. (2003) Acidovorax-like symbionts in the nephridia of earthworms. Environ Microbiol 5: 804809.
  • Schweitzer, B., Huber, I., Amann, R., Ludwig, W., and Simon, M. (2001) Alpha- and beta-proteobacteria control the consumption and release of amino acids on lake snow aggregates. Appl Environ Microbiol 67: 632645.
  • Sims, R.W., and Gerard, B.M. (1985) Earthworms, Synopses of the British Fauna No. 31. Leiden, the Netherlands: E. J. Brill/Dr. W. Backhuys.
  • Thompson, J., Gibson, T., Plewniak, F., Jeanmougin, F., and Higgins, D. (1997) The clustal_x windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25: 48764882.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Experimental procedures
  6. Acknowledgements
  7. References
  8. Supporting Information

Fig. S1. Symbiont phylogeny based on 16S rRNA gene sequences. Strict consensus tree from neighbour joining (NJ), maximum likelihood (ML) and Bayesian inference (BI). Branch lengths are based on BI. Full circles represent posterior probabilities above 90. Open circles represent posterior probabilities between 67 and 90.

Fig. S2. Earthworm phylogeny based on concatenated H3 and COI gene nucleotide sequences. Strict consensus tree from neighbour joining (NJ), maximum likelihood (ML) and Bayesian inference (BI). Branch lengths are based on BI. Numbers on branches are posterior probabilities.

Fig. S3. Symbiont phylogeny based on rpoB nucleotide gene sequences. Strict consensus tree from neighbour joining (NJ), maximum likelihood (ML) and Bayesian inference (BI). Branch lengths are based on BI. Numbers on branches are posterior probabilities.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

EMI_2084_sm_Fig_S1.pdf64KSupporting info item
EMI_2084_sm_Fig_S2.pdf49KSupporting info item
EMI_2084_sm_Fig_S3.pdf54KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.