Distinct effects of the nephridial symbionts Verminephrobacter and Candidatus Nephrothrix on reproduction and maturation of its earthworm host Eisenia andrei

Distinct effects of the nephridial symbionts Verminephrobacter and Candidatus Nephrothrix on... Abstract Verminephrobacter, the most common specific symbionts in the nephridia (excretory organs) of lumbricid earthworms, have been shown to improve reproduction of the garden earthworm Aporrectodea tuberculata under nutrient limitation. It is unknown how general this beneficial trait is in the Verminephrobacter–earthworm symbiosis, whether other nephridial symbionts also affect host fitness and what the mechanism of the fitness increase is. Here we report beneficial effects of Verminephrobacter and Candidatus Nephrothrix on life history traits of the compost worm Eisenia andrei, which in addition to these two symbionts also hosts Agromyces-like bacteria in its mixed nephridial community: while growth was identical between control, Verminephrobacter-free and aposymbiotic worms, control worms produced significantly more cocoons and offspring than both Verminephrobacter-free and aposymbiotic worms, confirming the reproductive benefit of Verminephrobacter in a second host with different ecology and feeding behavior. Furthermore, worms with Verminephrobacter and Ca. Nephrothrix, or with only Ca. Nephrothrix present, reached sexual maturity significantly earlier than aposymbiotic worms; this is the first evidence for a beneficial role of Ca. Nephrothrix in earthworms. Riboflavin content in cocoons and whole earthworms was unaffected by the presence or absence of nephridial symbionts, suggesting that nutritional supplementation with this vitamin does not play a major role in this symbiosis. host–microbe interaction, multi-partner symbiosis, fitness experiment, Eisenia andrei, nephridia INTRODUCTION Lumbricid earthworms carry in their excretory organs (nephridia) specific extracellular symbionts (Schramm et al.2003; Lund et al.2010a; Davidson, Powell and James 2013), with Verminephrobacter (Betaproteobacteria) as the most common and best-studied example (Lund, Kjeldsen and Schramm 2014). Verminephrobacter are species-specific (Lund et al.2010a), vertically transmitted via the cocoon (Davidson and Stahl 2006), and, in the common garden earthworm Aporrectodea tuberculata, increase host reproductive fitness under nutrient limitation, i.e. the presence of Verminephrobacter results in earlier sexual maturation and higher hatching success compared to aposymbiotic worms (Lund et al.2010b). It is currently unknown whether this beneficial effect is a general trait of the earthworm–Verminephrobacter symbiosis, and whether the additional bacterial symbionts that are present in most lumbricid earthworms also affects host fitness. In multipartner interactions, the functional organization of the different symbionts might enhance their host fitness by having complementary functions or by providing the same function in a cumulative or synergistic manner. This is thoroughly described within the endosymbionts of Hemipteran insects where in some cases multiple symbionts provide the host with essential amino acids and in other cases one of the symbionts provides amino acids while the other protects the host against pathogens (Douglas 2016). Additionally, one of the whitefly endosymbionts has also been suggested to be a metabolic parasite while the other symbiont is providing the host with essential amino acids (Ankrah, Luan and Douglas 2017). If and how the multiple earthworm nephridial symbionts affects host fitness is currently unknown. The common compost worms, Eisenia andrei and E. fetida, harbor mixed nephridial communities comprised of Verminephrobacter, Ca. Nephrothrix (Bacteroidetes) and an Agromyces-like (Microbacteriaceae) symbiont (Davidson et al.2010; Davidson, Powell and James 2013). Candidatus Nephrothrix is species-specific and has been detected in about half of all investigated lumbricid earthworm species (Davidson, Powell and James 2013; Møller, Lund and Schramm 2015), while Agromyces-like symbionts have only been reported in Eisenia (Davidson, Powell and James 2013). Like Verminephrobacter, both Ca. Nephrothrix and the Agromyces-like symbionts are vertically transmitted via the cocoon (Davidson et al.2010). Furthermore, by differential antibiotics treatment of the cocoons, either Verminephrobacter alone or all nephridial symbionts can be removed from worm cultures in the laboratory (Dulla et al.2012), resulting in an ideal model to test the differential effect of these two symbionts in Eisenia. The mechanism underlying Verminephrobacter’s fitness benefit in A. tuberculata is still unknown but considering that it is only evident under nutrient limitation, it is likely that the symbionts provide a nutritional supplement to their host. It has previously been hypothesized that the symbionts enhance earthworm nitrogen retention through excretion of proteolytic enzymes that degrade small proteins and peptides otherwise lost in the urine, thereby enabling the host to reabsorb the valuable amino acids (Padazis 1931). This hypothesis, however, fails to explain the symbionts’ effect on cocoon hatching success, as cocoon amino acid content was not affected by symbiont presence (Lund et al.2010b). Alternatively, the symbionts may provide essential vitamins to their host, a common trait in other symbioses (Douglas 2014; Martín et al.2014). When analyzing riboflavin content in the compost worm E. andrei, Sulik et al. (2012) found no difference between starved and well-fed worms and therefore hypothesized that the nephridial symbionts were the source of riboflavin for their host. The objectives of our study were thus (i) to investigate the fitness effect of Verminephrobacter compared to that of Ca. Nephrothrix and Agromyces-like symbionts in the compost worm E. andrei and (ii) to test the riboflavin supplementation hypothesis (Sulik et al.2012) as a mechanistic explanation for a possible beneficial effect. MATERIALS AND METHODS Earthworm cultures Adult individuals of the epigeic compost worm, E. andrei, were collected from a compost pile in Aarhus (Denmark) in Spring 2012. They were identified by morphological characteristics (Sims and Gerard 1985) and by sequencing the mitochondrially encoded cytochrome c oxidase subunit I gene (Møller, Lund and Schramm 2015, GenBank accession numbers KP420568-KP420571). The worms were kept at 20°C in 1-L plastic pots containing soil supplemented with grounded cow dung as a food source. Cow dung was added every 3 weeks and the bedding soil was exchanged every 6 weeks. Both soil and cow dung were dried at 80°C for 24 h prior to use in order to kill indigenous soil fauna. Verminephrobacter-free and aposymbiotic worm lines were established by differential antibiotic treatment of young cocoons (Davidson and Stahl 2006). Freshly laid cocoons were collected every week by manual sorting of the bedding soil, washed thoroughly in tap water and incubated at 20°C for 10 days on filter paper soaked in antibiotics. Filter paper and antibiotics were replaced daily. The Verminephrobacter-free culture was established by treating the cocoons with kanamycin sulfate (150 μg/mL), and the aposymbiotic worm culture was established by treating the cocoons with both kanamycin sulfate (150 μg/mL) and erythromycin (150 μg/mL). Hatchlings were transferred to pots with soil and maintained as described before. Only worms from the second generation, which had never been exposed to antibiotics, were used in the fitness experiments to avoid any possible impact caused by the treatment. Symbiont detection in the earthworm cultures The presence or absence of the three symbionts in the worm cultures was checked by FISH as previously described (Lund et al.2010a). In the first generation, three individuals from each treatment were investigated, whereas ten individuals from each treatment were investigated in the second generation. Briefly, adult individuals from each worm line were dissected, body wall tissue with nephridia still attached were fixed in 4% paraformaldehyde and hybridized. Finally, the hybridized nephridia were picked off the body wall with forceps and mounted for epifluorescence microscopy. A Cy3-labeled probe specific for one of the three symbionts (LSB145 for Verminephrobacter and some Acidovorax (Schweitzer et al.2001): FLX226 for Ca. Nephrothrix (Møller, Lund and Schramm 2015); LEIF841 for Agromyces-like earthworms symbionts) was combined with a Cy5-labeled bacterial probe (EUB338-II-III; Daims et al.1999); probe NON (Manz et al.1992) was included as negative control. Probe sequences and formamide concentrations used during hybridization are given in Table S1 (Supporting Information). The presence of Agromyces-like symbionts was also assessed by specific PCR in hatchlings from the fitness experiment. Five hatchlings from each worm line were homogenized by bead beating in 20 μL dH2O, and the homogenate was used for direct PCR with the TerraTM PCR Direct Polymerase Mix (Takara) with the specific primers EfA243F (Davidson et al.2010) and Leif842R. Thermal cycling conditions and primer sequences are presented in Table S1. Comparative fitness experiment Growth, time to sexual maturity, cocoon production and hatching success of E. andrei were investigated under nutrient-limiting conditions for the three earthworm lines. Cocoons were collected during a period of 2 months from each of the three earthworm lines and stored in petri dishes with moistened filter paper at 4°C to halt embryonic development. When a number of 20–30 cocoons per group was reached, they were moved to 20°C to induce hatching, which occurred over a period of 1 month. Hatchlings were kept in petri dishes with moist filter paper at 4°C to prevent growth until the experiment started. Per group, 25 hatchlings were transferred to pots with 10 g moist soil, kept at 20°C for the entire experiment, and fed every 3 weeks, when both bedding soil and feed were replaced (Table S2, Supporting Information). The initial straw diet was substituted with cow dung after 21 weeks to improve worm growth. Straw and cow dung were dried at 80°C for 24 h, thoroughly mixed with dried soil and tap water was added to a water content of 20% dry weight (dw) of the mixture. Total nitrogen and total organic carbon were 0.23 mmol N g−1 and 36.83 mmol C g−1 in straw, 1.47 mmol N g−1 and 33.42 mmol C g−1 in cow dung, and 0.13 mmol N g−1 and 2.36 mmol C g−1 in soil, resulting in C/N ratios of 163.8, 22.7 and 18.5 for straw, cow dung and soil, respectively. Initially, worms received food equal to 0.25% dw of the soil, and this was adjusted as the worms grew to a final amount of 0.4% dw of soil; likewise, the soil volume was increased successively to 100 g; for a detailed feeding scheme, see Table S2 (Supporting Information). Every 3 weeks, the worms were washed in tap water, blotted dry on filter paper and weighted. Time of sexual maturity, determined by visible tubercula pubertatis, was recorded; when all worms were sexually mature (after 39 weeks), they were mated in pairs, which were kept together until the end of the experiment. Cocoons were hand-picked every 10–11 days, placed individually in petri dishes with moist filter paper and incubated at 20°C until hatching. The number of hatchlings per cocoon was recorded, and cocoons that did not hatch within 2 months after collection were recorded as ‘not hatched’. Riboflavin measurements Riboflavin content was analyzed in six whole worms per group and in three pools of two to three cocoons (to reach a total weight of >0.2 mg). Worms and cocoons were washed in tap water, blotted dry, weighted, placed individually in microcentrifuge tubes, flash-frozen in liquid nitrogen and stored at –80°C until further analysis. For extraction of riboflavin from whole worms, 1 mL of extraction solution (methanol/water, 1/1 v/v) was added to each frozen worm in brown tubes (riboflavin is light sensitive) containing FastPrep-Lysing Matrix D (1.4 mm ceramic spheres; MoBio Laboratories). Worms were homogenized by bead beating in two cycles of 40 s at 6.0 m s−1 with cooling at 0°C for 2 min between the cycles. After centrifugation at 13 000 rpm at 0°C for 5 min, the supernatant was transferred into a new tube. The extraction was repeated twice and supernatants of the same worm were pooled and stored at –80°C. Extraction from cocoons was identical except that cocoons were opened with forceps, the number of bead beating cycles was increased to 6, and only 0.5 mL and 0.25 mL extraction solution was used in the first and second round of extraction, respectively. For detection and quantification of riboflavin, a modified RP-HPLC was performed using a Dionex® HPLC System (p680 HPLC pump; ASI-100 autosampler) coupled to a fluorescence detector (RF-10A XL Schimadzu®) (Andrés-Lacueva, Mattivi and Tonon 1998). Chromatographic separation was performed using a Prontosil® Eurobond C18 column (5 μm, 125 mm × 4 mm). The flow rate was 0.6 mL min−1 and the total data acquisition running time was 15 min. The column was maintained at 25°C, with an injection volume of 10 μL and the HPLC parameters were the following: solvent for mobile phase A was 50 mM NaH2PO4, pH = 4.5 and solvent for mobile phase B was acetonitrile. The gradient elution started with 15% B (0–1 min), then 15%–50% B (1–7 min), followed by 50–95% (7–8 min) and 95% (8–10 min), return to 15% B (10–11 min) and finally 15% B until end of run. The fluorescence detector was set at 445 nm for the excitation wavelength and 525 nm for the emission wavelength. For absolute quantification, a seven-point calibration between 2.5 and 500 μM was generated. Statistical tests The time of sexual maturity was fitted to a flexible Richard's growth model. The number of sexually mature worms at time t was assumed to be binomially distributed Bin[n, p(t)], where n was the number of worms and p(t) is the probability of having reached sexual maturity at time t. The probability parameter p(t) was modeled by a four-parameter Richard's growth model (Seber and Wild 1989), with only three free parameters as the parameters that measure the asymptotic value was fixed to 1; the point of inflection was parameterized by two parameters: one parameter that measured the timing of the point of inflection (x axis) and a second parameter that measured at what proportion of sexually mature worms the curve started to become concave (y axis). The third parameter measured the slope of the increase in the number of sexually mature worms. Statistical inferences about the time of sexual maturity were made using likelihood ratio tests. Richard's growth models were also used to describe the cumulative cocoon and hatchling productions; the number of cocoons or hatchlings was assumed to be Poisson distributed, where lambda (t) is the mean number of cocoons or hatchlings at time t. Lambda (t) was modeled as a four-parameter Richard's growth model as described above, except all parameters were free. Statistical inferences about differences in cumulative cocoon or hatchling production were made using likelihood ratio tests. Fisher's exact test was used to test if there was a significant difference in number of cocoon-producing couples between the three treatments. Potential differences in cocoon hatching success were tested using Excel Chitest. The correlation between number of cocoons produced per couple and the weight of each couple before (week 39) and after (week 66) cocoon production was assessed using the Pearson's product-moment correlation coefficient in R commander (Fox 2005). Difference in worm riboflavin content was analyzed using the non-parametric Wilcoxon signed rank test in R commander (Fox 2005). RESULTS AND DISCUSSION Establishment and validation of differential earthworm lines The presence of Verminephrobacter, Ca. Nephrothrix and (sporadically, see below) the Agromyces-like symbionts was confirmed by fluorescence in situ hybridization (FISH, Fig. S1, Supporting Information) in the E. andrei control culture, which was started from locally collected compost worms. Kanamycin treatment of E. andrei cocoons resulted in a Verminephrobacter-free worm line (as proven by FISH, Fig. S1, Supporting Information), while treatment with both kanamycin and erythromycin resulted in a fully aposymbiotic worm line (as proven by FISH). The Agromyces-like symbiont was detected by FISH and specific PCR in both the control and the Verminephrobacter-free worm line but not in all individuals; PCR confirmed it in two out of five control individuals, and three out of five Verminephrobacter-free individuals. FISH results indicated a very low abundance (i.e. a few cells per nephridium) and it was not detected in all individuals. These observations are consistent with Davidson et al. (2010), who also reported sporadic presence and low numbers of these symbionts in the nephridia. At such low and sporadic abundance, we consider it unlikely that the Agromyces-like symbiont contributes to the fitness effects measured in our experiment, which we therefore ascribe to the Verminephrobacter and Ca. Nephrothrix symbionts. Nephridial symbionts affect sexual maturation but not growth of their E. andrei host Growth and reproduction of the three worm lines (control, Verminephrobacter-free, aposymbiotic) was compared over a complete E. andrei life cycle under nutrient limitation; this condition was chosen because the fitness benefit of Verminephrobacter in the common garden worm A. tuberculata was only apparent under nutrient limitation (Lund et al.2010b). Consistent with this earlier study (Lund et al.2010b), growth of the worms during the first 39 weeks (measured as weight gain) was not affected by the removal of Verminephrobacter or by the removal of all three symbionts (Fig. 1). Resource limitation was apparent by the drastically prolonged generation times: while well-fed E. andrei can commonly grow to adulthood and develop clitella in only 5 weeks (Reinecke and Viljoen 1991; Domínguez, Briones and Mato 1997), visible tubercula pubertatis and thus sexual maturity in our experiment was first detected after 27–39 weeks (Fig. 2). The time of sexual maturation was slightly delayed but not significantly different from the control in the Verminephrobacter-free E. andrei line (Fig. 2); this is different from A. tuberculata, where removal of Verminephrobacter significantly delayed sexual maturation (Lund et al.2010b). Apparently, Ca. Nephrothrix (and possibly the Agromyces-like symbiont) can partially compensate for the loss of Verminephrobacter, as only the complete removal of the symbionts had a significant effect on sexual maturation (Richard's growth model, P < 0.0001; Fig. 2). Thus, both Verminephrobacter and Ca. Nephrothrix (and the Agromyces-like symbiont) affect the timing of sexual maturation in a cumulative or synergistic manner. Figure 1. View largeDownload slide Growth of E. andrei during 39 weeks and the final weight of the worms in week 66 (after cocoon production). Bars represent averages with standard deviations (n = 20) for control, Verminephrobacter-free and aposymbiotic worms, according to legend. Figure 1. View largeDownload slide Growth of E. andrei during 39 weeks and the final weight of the worms in week 66 (after cocoon production). Bars represent averages with standard deviations (n = 20) for control, Verminephrobacter-free and aposymbiotic worms, according to legend. Figure 2. View largeDownload slide Number of sexually mature worms over time in each group. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 2. View largeDownload slide Number of sexually mature worms over time in each group. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Verminephrobacter but not Ca. Nephrothrix benefit cocoon and offspring production The cumulative cocoon and hatchling production is shown in Fig. 3; the control worms produced significantly more cocoons and hatchlings than both the Verminephrobacter-free and the aposymbiotic worms (Richard's growth model, P < 0.0001: Fig. 3). The hatching success of the cocoons ranged from 58% to 81% but was not significantly different between the treatments (Chi test, P = 0.155; Fig. 4). These combined results indicate that Ca. Nephrothrix had no effect (neither beneficial nor detrimental) on cocoon production or hatching success, while Verminephrobacter is beneficial for E. andrei reproduction by increasing cocoon and thereby hatchling numbers. Verminephrobacter also benefitted reproduction in A. tuberculata; however, here cocoon hatching success (57% in symbiotic vs 25% in aposymbiotic worms) and not cocoon production was affected (Lund et al.2010b). It is unknown why Verminephrobacter affects cocoon production and hatching success differently in the two earthworm species. One explanation could be that hatching success in E. andrei is affected by the number of times the worm has mated (Porto, Velando and Domínguez 2012); if a worm only mated once, the hatching success was 59 ± 11% (same range as this study, Fig. 4), whereas worms mated two to six times had a cocoon hatching success of 80 ± 5%. We did not control the number of matings per worm, which may influence, and thereby obscure, a potential effect of Verminephrobacter on cocoon hatching success in E. andrei. It has, to our knowledge, never been investigated if the number of mating events also influences hatching success in A. tuberculata. Figure 3. View largeDownload slide Cumulative production of cocoons (A) and hatchlings (B) in control, Verminephrobacter-free and aposymbiotic worms. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 3. View largeDownload slide Cumulative production of cocoons (A) and hatchlings (B) in control, Verminephrobacter-free and aposymbiotic worms. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 4. View largeDownload slide Cocoon production and hatching success. Gray and white bars show number of hatched and non-hatched cocoons, respectively. Numbers above bars show hatching success in %. There was no significant difference in hatching success between the three treatments (Chi test, P = 0.155). Figure 4. View largeDownload slide Cocoon production and hatching success. Gray and white bars show number of hatched and non-hatched cocoons, respectively. Numbers above bars show hatching success in %. There was no significant difference in hatching success between the three treatments (Chi test, P = 0.155). Even though the growth of E. andrei was not directly affected by the symbionts (Fig. 1), worms in the control group were able to reproduce at a smaller size than both Verminephrobacter-free and aposymbiotic worms (Fig. 5). If the Verminephrobacter-free and aposymbiotic worms were below a weight threshold of about 0.2 g in week 39, they never produced cocoons, whereas the control worms below this threshold produced one to four cocoons. Control worms consistently produced more cocoons even though they had a similar weight as the Verminephrobacter-free and aposymbiotic worms. Overall, the weight of the worms at the time of mating (week 39) positively correlated with the number of cocoons the couples would produce during the following 27 weeks (Pearson's r = 0.6923, 0.8480 and 0.7417 for the control, Verminephrobacter-free and aposymbiotic groups, respectively, P < 0.005; Fig. 5). The ability of control worms to produce more offspring at a smaller size suggests that Verminephrobacter (but not Ca. Nephrothrix) provide a nutritional supplement (e.g. vitamins) to their host, which allows them to invest in offspring early on. Figure 5. View largeDownload slide Number of cocoons produced per couple as a function of the weight of each worm in week 39. Worms were mated in pairs of two and each couple produced x cocoons. If several couples from the same treatment produced the same number of cocoons, the couples can be differentiated according to the color; two individuals from the same couple have the same color. There is a significant correlation between the weight of the worms and how many cocoons they will produce during the following 27 weeks (control; blue line, Pearson r = 0.6923. Verminephrobacter-free; red line, Pearson r = 0.8480. Aposymbiotic; green line, Pearson r = 0.7417. P < 0.005). If the Verminephrobacter-free and aposymbiotic groups are below a weight threshold of about 0.2 g in week 39 (vertical gray line), they never produce cocoons, whereas the control worms below this threshold produce one to four cocoons. Figure 5. View largeDownload slide Number of cocoons produced per couple as a function of the weight of each worm in week 39. Worms were mated in pairs of two and each couple produced x cocoons. If several couples from the same treatment produced the same number of cocoons, the couples can be differentiated according to the color; two individuals from the same couple have the same color. There is a significant correlation between the weight of the worms and how many cocoons they will produce during the following 27 weeks (control; blue line, Pearson r = 0.6923. Verminephrobacter-free; red line, Pearson r = 0.8480. Aposymbiotic; green line, Pearson r = 0.7417. P < 0.005). If the Verminephrobacter-free and aposymbiotic groups are below a weight threshold of about 0.2 g in week 39 (vertical gray line), they never produce cocoons, whereas the control worms below this threshold produce one to four cocoons. Verminephrobacter symbionts do not alter the riboflavin content of their hosts Further indications for a symbiotic nutrient supplementation that in part compensates for a poor diet are the observation that (i) nutrient limitation was required to detect a fitness effect of Verminephrobacter in A. tuberculata (Lund et al.2010b); (ii) aposymbiotic stock cultures of both E. andrei and A. tuberculata are reproducing normally under nutrient-rich conditions in the lab; and (iii) the fitness effect of the symbionts was similar in both worm species. One possibility for such a supplementation is riboflavin, as recently hypothesized for E. andrei (Sulik et al.2012). We found that the content of riboflavin in cocoons collected from worm stock cultures (not starved) did not differ between the control, Verminephrobacter-free or aposymbiotic worms (Wilcoxon signed-rank test, P = 0.25) (Fig. S2, Supporting Information). Similarly, there were no differences in riboflavin content between worm lines at the end of the fitness experiment (Wilcoxon signed-rank test, P > 0.6, Fig. S3, Supporting Information). Consequently, as Verminephrobacter does not affect earthworm riboflavin content, it is unlikely that it provides the host with this vitamin. CONCLUSION AND FUTURE PERSPECTIVES The collective findings from this study and Lund et al. (2010b) show that in both the compost worm E. andrei and the garden worm A. tuberculata the Verminephrobacter symbionts enhance host reproduction, and that both worm species reach sexual maturity earlier when all of their native symbionts are present. Intriguingly, these beneficial effects appear conserved between host species despite the different ecology and feeding behavior of these earthworms. As these effects are seen under nutrient-limiting conditions, nutritional supplementation appears an obvious explanation. Yet neither the original amino acid recycling proposal (Pandazis 1931) nor the more recent riboflavin hypothesis (Sulik et al.2012) could be confirmed in our comparative fitness studies (Lund et al.2010b; this study). Since the nutritional status of the worm is still likely to be a determinant for the onset of sexual maturity and reproductive success, the question thus remains: What compound(s) do the symbionts provide for their host? Or do they rather detoxify waste products that otherwise weaken the host? Alternatively, the symbionts may also directly manipulate the host reproductive hormone system to ensure early and prolific reproduction, thereby benefitting their own transfer to the next generation. Interestingly, the receptor for the hormone anneotcin, which induces egg-laying behavior in earthworms (Oumi et al.1996), is produced in the nephridia of the clitellum region (Kawada et al.2004), thereby co-localizing with the nephridial symbionts. Manipulation of host reproductive hormones, however, does not explain why the fitness effect is seen under nutrient limitation. Thus, the molecular mechanism behind the fitness benefit of nephridial earthworm symbionts is still to be discovered. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. Acknowledgements We thank Annemarie Højmark, Susanne Nielsen and Karina Bomholt Oest for excellent technical assistance and Martin Holmstrup for good advice and fruitful discussions on earthworm ecology. FUNDING This work was funded by the EU Marie Curie Initial Training Network (ITN) Symbiomics (Contract Number 264774), the Aarhus Institute of Advanced Studies (AIAS), and the Danish Council for Independent Research | Natural Sciences (FNU). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. Conflict of interest. None declared. REFERENCES Andrés-Lacueva C, Mattivi F, Tonon D. Determination of riboflavin, flavin mononucleotide and flavin–adenine dinucleotide in wine and other beverages by high performance liquid chromatography with fluorescence detection. J Chromatogr A  1998; 823: 355– 63. Google Scholar CrossRef Search ADS PubMed  Ankrah NYD, Luan J, Douglas AE. Cooperative metabolism in a three-partner insect-bacterial symbiosis revealed by metabolic modeling. J Bacteriol  2017, doi:10.1128/JB.00872-16. Daims H, Brühl A, Amann R et al.   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  1999; 22: 434– 44. Google Scholar CrossRef Search ADS PubMed  Davidson SK, Powell RJ, James S. A global survey of the bacteria within earthworm nephridia. Mol Phylogenet Evol  2013; 67: 188– 200. Google Scholar CrossRef Search ADS PubMed  Davidson SK, Powell RJ, Stahl DA. Transmission of a bacterial consortium in Eisenia fetida egg capsules. Environ Microbiol  2010; 12: 2277– 88. Google Scholar PubMed  Davidson SK, Stahl D. Transmission of nephridial bacteria of the earthworm Eisenia fetida. Appl Environ Microb  2006; 72: 769– 75. Google Scholar CrossRef Search ADS   Domínguez J, Briones M, Mato S. Effect on the diet on growth and reproduction of Eisenia andrei (Oligochaeta, Lumbricidae). Pedobiologia  1997; 41: 566– 76. Douglas AE. The molecular basis of bacterial-insect symbiosis. J Mol Biol  2014; 426: 3830– 7. Google Scholar CrossRef Search ADS PubMed  Douglas AE. How multi-partner endosymbiosesfunction. Nat Rev Microbiol  2016; 14: 731– 43. Google Scholar CrossRef Search ADS PubMed  Dulla G, Go R, Stahl D et al.   Verminephrobacter eiseniae type IV pili and flagella are required to colonize earthworm nephridia. ISME J  2012; 6: 1166– 75. Google Scholar CrossRef Search ADS PubMed  Fox J. The R commander: A basic-statistics graphical user interface to R. J Stat Softw  2005; 14: 1– 42. Kawada T, Kanda A, Minakata H et al.   Identification of a novel receptor for an invertebrate oxytocin/vasopressin superfamily peptide: molecular and functional evolution of the oxytocin/vasopressin superfamily. Biochem J  2004; 382: 231– 7. Google Scholar CrossRef Search ADS PubMed  Lund MB, Davidson SK, Holmstrup M et al.   Diversity and host specificity of the Verminephrobacter-earthworm symbiosis. Environ Microbiol  2010a; 12: 2142– 51. Lund MB, Holmstrup M, Lomstein B et al.   Beneficial effect of Verminephrobacter nephridial symbionts on the fitness of the earthworm Aporrectodea tuberculata. Appl Environ Microb  2010b; 76: 4738– 43. Google Scholar CrossRef Search ADS   Lund MB, Kjeldsen KU, Schramm A. The earthworm-Verminephrobacter symbiosis: an emerging experimental system to study extracellular symbiosis. Front Microbiol  2014; 5: 128. Google Scholar CrossRef Search ADS PubMed  Manz W, Amann R, Ludwig W et al.   Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria—problems and solutions. Syst Appl Microbiol  1992; 15: 593– 600 Google Scholar CrossRef Search ADS   Martín R, Miquel S, Ulmer J et al.   Gut ecosystem: how microbes help us. Benef Microbes  2014; 5: 219– 33. Google Scholar CrossRef Search ADS PubMed  Møller P, Lund MB, Schramm A. Evolution of the tripartite symbiosis between earthworms, Verminephrobacter and Flexibacter-like bacteria. Front Microbiol  2015; 6: 529 Google Scholar PubMed  Oumi T, Ukena K, Matsushima O et al.   Annetocin, an annelid oxytocin-related peptide, induces egg-laying behavior in the earthworm, Eisenia foetida. J Exp Zool  1996; 276: 151– 6. Google Scholar CrossRef Search ADS PubMed  Pandazis G. Zur Frage der Bakteriensymbiose bei Oligochäten. Zentlbl Bakteriol  1931; 120: 440– 53. Porto PG, Velando A, Domínguez J. Multiple mating increases cocoon hatching success in the earthworm Eisenia andrei (Oligochaeta: Lumbricidae). Biol J Linnean Soc  2012; 107: 175– 81. Google Scholar CrossRef Search ADS   Reinecke AJ, Viljoen SA. A comparison of the biology of Eisenia fetida and Eisenia andrei (Oligochaeta). Biol Fertil Soils  1991; 11: 295– 300. Google Scholar CrossRef Search ADS   Schramm A, Davidson SK, Dodsworth J et al.   Acidovorax-like symbionts in the nephridia of earthworms. Environ Microbiol  2003; 5: 804– 9. Google Scholar CrossRef Search ADS PubMed  Schweitzer B, Huber I, Amann R et al.   α-and β-Proteobacteria control the consumption and release of amino acids on lake snow aggregates. Appl Environ Microb  2001; 67: 632– 45. Google Scholar CrossRef Search ADS   Seber GAF, Wild CJ. Nonlinear Regression . New York: John Wiley & Sons, Inc., 1989. Google Scholar CrossRef Search ADS   Sims RW, Gerard BM. Earthworms . Leiden: J. Brill/Dr. W. Backhuys, 1985. Sulik P, Klimek M, Talik P et al.   Searching for external sources of the riboflavin stored in earthworm eleocytes. Invertebr Surv J  2012; 9: 169– 77. © FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png FEMS Microbiology Ecology Oxford University Press

Distinct effects of the nephridial symbionts Verminephrobacter and Candidatus Nephrothrix on reproduction and maturation of its earthworm host Eisenia andrei

Loading next page...
 
/lp/ou_press/distinct-effects-of-the-nephridial-symbionts-verminephrobacter-and-nwiF06he0X
Publisher
Blackwell
Copyright
© FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0168-6496
eISSN
1574-6941
D.O.I.
10.1093/femsec/fix178
Publisher site
See Article on Publisher Site

Abstract

Abstract Verminephrobacter, the most common specific symbionts in the nephridia (excretory organs) of lumbricid earthworms, have been shown to improve reproduction of the garden earthworm Aporrectodea tuberculata under nutrient limitation. It is unknown how general this beneficial trait is in the Verminephrobacter–earthworm symbiosis, whether other nephridial symbionts also affect host fitness and what the mechanism of the fitness increase is. Here we report beneficial effects of Verminephrobacter and Candidatus Nephrothrix on life history traits of the compost worm Eisenia andrei, which in addition to these two symbionts also hosts Agromyces-like bacteria in its mixed nephridial community: while growth was identical between control, Verminephrobacter-free and aposymbiotic worms, control worms produced significantly more cocoons and offspring than both Verminephrobacter-free and aposymbiotic worms, confirming the reproductive benefit of Verminephrobacter in a second host with different ecology and feeding behavior. Furthermore, worms with Verminephrobacter and Ca. Nephrothrix, or with only Ca. Nephrothrix present, reached sexual maturity significantly earlier than aposymbiotic worms; this is the first evidence for a beneficial role of Ca. Nephrothrix in earthworms. Riboflavin content in cocoons and whole earthworms was unaffected by the presence or absence of nephridial symbionts, suggesting that nutritional supplementation with this vitamin does not play a major role in this symbiosis. host–microbe interaction, multi-partner symbiosis, fitness experiment, Eisenia andrei, nephridia INTRODUCTION Lumbricid earthworms carry in their excretory organs (nephridia) specific extracellular symbionts (Schramm et al.2003; Lund et al.2010a; Davidson, Powell and James 2013), with Verminephrobacter (Betaproteobacteria) as the most common and best-studied example (Lund, Kjeldsen and Schramm 2014). Verminephrobacter are species-specific (Lund et al.2010a), vertically transmitted via the cocoon (Davidson and Stahl 2006), and, in the common garden earthworm Aporrectodea tuberculata, increase host reproductive fitness under nutrient limitation, i.e. the presence of Verminephrobacter results in earlier sexual maturation and higher hatching success compared to aposymbiotic worms (Lund et al.2010b). It is currently unknown whether this beneficial effect is a general trait of the earthworm–Verminephrobacter symbiosis, and whether the additional bacterial symbionts that are present in most lumbricid earthworms also affects host fitness. In multipartner interactions, the functional organization of the different symbionts might enhance their host fitness by having complementary functions or by providing the same function in a cumulative or synergistic manner. This is thoroughly described within the endosymbionts of Hemipteran insects where in some cases multiple symbionts provide the host with essential amino acids and in other cases one of the symbionts provides amino acids while the other protects the host against pathogens (Douglas 2016). Additionally, one of the whitefly endosymbionts has also been suggested to be a metabolic parasite while the other symbiont is providing the host with essential amino acids (Ankrah, Luan and Douglas 2017). If and how the multiple earthworm nephridial symbionts affects host fitness is currently unknown. The common compost worms, Eisenia andrei and E. fetida, harbor mixed nephridial communities comprised of Verminephrobacter, Ca. Nephrothrix (Bacteroidetes) and an Agromyces-like (Microbacteriaceae) symbiont (Davidson et al.2010; Davidson, Powell and James 2013). Candidatus Nephrothrix is species-specific and has been detected in about half of all investigated lumbricid earthworm species (Davidson, Powell and James 2013; Møller, Lund and Schramm 2015), while Agromyces-like symbionts have only been reported in Eisenia (Davidson, Powell and James 2013). Like Verminephrobacter, both Ca. Nephrothrix and the Agromyces-like symbionts are vertically transmitted via the cocoon (Davidson et al.2010). Furthermore, by differential antibiotics treatment of the cocoons, either Verminephrobacter alone or all nephridial symbionts can be removed from worm cultures in the laboratory (Dulla et al.2012), resulting in an ideal model to test the differential effect of these two symbionts in Eisenia. The mechanism underlying Verminephrobacter’s fitness benefit in A. tuberculata is still unknown but considering that it is only evident under nutrient limitation, it is likely that the symbionts provide a nutritional supplement to their host. It has previously been hypothesized that the symbionts enhance earthworm nitrogen retention through excretion of proteolytic enzymes that degrade small proteins and peptides otherwise lost in the urine, thereby enabling the host to reabsorb the valuable amino acids (Padazis 1931). This hypothesis, however, fails to explain the symbionts’ effect on cocoon hatching success, as cocoon amino acid content was not affected by symbiont presence (Lund et al.2010b). Alternatively, the symbionts may provide essential vitamins to their host, a common trait in other symbioses (Douglas 2014; Martín et al.2014). When analyzing riboflavin content in the compost worm E. andrei, Sulik et al. (2012) found no difference between starved and well-fed worms and therefore hypothesized that the nephridial symbionts were the source of riboflavin for their host. The objectives of our study were thus (i) to investigate the fitness effect of Verminephrobacter compared to that of Ca. Nephrothrix and Agromyces-like symbionts in the compost worm E. andrei and (ii) to test the riboflavin supplementation hypothesis (Sulik et al.2012) as a mechanistic explanation for a possible beneficial effect. MATERIALS AND METHODS Earthworm cultures Adult individuals of the epigeic compost worm, E. andrei, were collected from a compost pile in Aarhus (Denmark) in Spring 2012. They were identified by morphological characteristics (Sims and Gerard 1985) and by sequencing the mitochondrially encoded cytochrome c oxidase subunit I gene (Møller, Lund and Schramm 2015, GenBank accession numbers KP420568-KP420571). The worms were kept at 20°C in 1-L plastic pots containing soil supplemented with grounded cow dung as a food source. Cow dung was added every 3 weeks and the bedding soil was exchanged every 6 weeks. Both soil and cow dung were dried at 80°C for 24 h prior to use in order to kill indigenous soil fauna. Verminephrobacter-free and aposymbiotic worm lines were established by differential antibiotic treatment of young cocoons (Davidson and Stahl 2006). Freshly laid cocoons were collected every week by manual sorting of the bedding soil, washed thoroughly in tap water and incubated at 20°C for 10 days on filter paper soaked in antibiotics. Filter paper and antibiotics were replaced daily. The Verminephrobacter-free culture was established by treating the cocoons with kanamycin sulfate (150 μg/mL), and the aposymbiotic worm culture was established by treating the cocoons with both kanamycin sulfate (150 μg/mL) and erythromycin (150 μg/mL). Hatchlings were transferred to pots with soil and maintained as described before. Only worms from the second generation, which had never been exposed to antibiotics, were used in the fitness experiments to avoid any possible impact caused by the treatment. Symbiont detection in the earthworm cultures The presence or absence of the three symbionts in the worm cultures was checked by FISH as previously described (Lund et al.2010a). In the first generation, three individuals from each treatment were investigated, whereas ten individuals from each treatment were investigated in the second generation. Briefly, adult individuals from each worm line were dissected, body wall tissue with nephridia still attached were fixed in 4% paraformaldehyde and hybridized. Finally, the hybridized nephridia were picked off the body wall with forceps and mounted for epifluorescence microscopy. A Cy3-labeled probe specific for one of the three symbionts (LSB145 for Verminephrobacter and some Acidovorax (Schweitzer et al.2001): FLX226 for Ca. Nephrothrix (Møller, Lund and Schramm 2015); LEIF841 for Agromyces-like earthworms symbionts) was combined with a Cy5-labeled bacterial probe (EUB338-II-III; Daims et al.1999); probe NON (Manz et al.1992) was included as negative control. Probe sequences and formamide concentrations used during hybridization are given in Table S1 (Supporting Information). The presence of Agromyces-like symbionts was also assessed by specific PCR in hatchlings from the fitness experiment. Five hatchlings from each worm line were homogenized by bead beating in 20 μL dH2O, and the homogenate was used for direct PCR with the TerraTM PCR Direct Polymerase Mix (Takara) with the specific primers EfA243F (Davidson et al.2010) and Leif842R. Thermal cycling conditions and primer sequences are presented in Table S1. Comparative fitness experiment Growth, time to sexual maturity, cocoon production and hatching success of E. andrei were investigated under nutrient-limiting conditions for the three earthworm lines. Cocoons were collected during a period of 2 months from each of the three earthworm lines and stored in petri dishes with moistened filter paper at 4°C to halt embryonic development. When a number of 20–30 cocoons per group was reached, they were moved to 20°C to induce hatching, which occurred over a period of 1 month. Hatchlings were kept in petri dishes with moist filter paper at 4°C to prevent growth until the experiment started. Per group, 25 hatchlings were transferred to pots with 10 g moist soil, kept at 20°C for the entire experiment, and fed every 3 weeks, when both bedding soil and feed were replaced (Table S2, Supporting Information). The initial straw diet was substituted with cow dung after 21 weeks to improve worm growth. Straw and cow dung were dried at 80°C for 24 h, thoroughly mixed with dried soil and tap water was added to a water content of 20% dry weight (dw) of the mixture. Total nitrogen and total organic carbon were 0.23 mmol N g−1 and 36.83 mmol C g−1 in straw, 1.47 mmol N g−1 and 33.42 mmol C g−1 in cow dung, and 0.13 mmol N g−1 and 2.36 mmol C g−1 in soil, resulting in C/N ratios of 163.8, 22.7 and 18.5 for straw, cow dung and soil, respectively. Initially, worms received food equal to 0.25% dw of the soil, and this was adjusted as the worms grew to a final amount of 0.4% dw of soil; likewise, the soil volume was increased successively to 100 g; for a detailed feeding scheme, see Table S2 (Supporting Information). Every 3 weeks, the worms were washed in tap water, blotted dry on filter paper and weighted. Time of sexual maturity, determined by visible tubercula pubertatis, was recorded; when all worms were sexually mature (after 39 weeks), they were mated in pairs, which were kept together until the end of the experiment. Cocoons were hand-picked every 10–11 days, placed individually in petri dishes with moist filter paper and incubated at 20°C until hatching. The number of hatchlings per cocoon was recorded, and cocoons that did not hatch within 2 months after collection were recorded as ‘not hatched’. Riboflavin measurements Riboflavin content was analyzed in six whole worms per group and in three pools of two to three cocoons (to reach a total weight of >0.2 mg). Worms and cocoons were washed in tap water, blotted dry, weighted, placed individually in microcentrifuge tubes, flash-frozen in liquid nitrogen and stored at –80°C until further analysis. For extraction of riboflavin from whole worms, 1 mL of extraction solution (methanol/water, 1/1 v/v) was added to each frozen worm in brown tubes (riboflavin is light sensitive) containing FastPrep-Lysing Matrix D (1.4 mm ceramic spheres; MoBio Laboratories). Worms were homogenized by bead beating in two cycles of 40 s at 6.0 m s−1 with cooling at 0°C for 2 min between the cycles. After centrifugation at 13 000 rpm at 0°C for 5 min, the supernatant was transferred into a new tube. The extraction was repeated twice and supernatants of the same worm were pooled and stored at –80°C. Extraction from cocoons was identical except that cocoons were opened with forceps, the number of bead beating cycles was increased to 6, and only 0.5 mL and 0.25 mL extraction solution was used in the first and second round of extraction, respectively. For detection and quantification of riboflavin, a modified RP-HPLC was performed using a Dionex® HPLC System (p680 HPLC pump; ASI-100 autosampler) coupled to a fluorescence detector (RF-10A XL Schimadzu®) (Andrés-Lacueva, Mattivi and Tonon 1998). Chromatographic separation was performed using a Prontosil® Eurobond C18 column (5 μm, 125 mm × 4 mm). The flow rate was 0.6 mL min−1 and the total data acquisition running time was 15 min. The column was maintained at 25°C, with an injection volume of 10 μL and the HPLC parameters were the following: solvent for mobile phase A was 50 mM NaH2PO4, pH = 4.5 and solvent for mobile phase B was acetonitrile. The gradient elution started with 15% B (0–1 min), then 15%–50% B (1–7 min), followed by 50–95% (7–8 min) and 95% (8–10 min), return to 15% B (10–11 min) and finally 15% B until end of run. The fluorescence detector was set at 445 nm for the excitation wavelength and 525 nm for the emission wavelength. For absolute quantification, a seven-point calibration between 2.5 and 500 μM was generated. Statistical tests The time of sexual maturity was fitted to a flexible Richard's growth model. The number of sexually mature worms at time t was assumed to be binomially distributed Bin[n, p(t)], where n was the number of worms and p(t) is the probability of having reached sexual maturity at time t. The probability parameter p(t) was modeled by a four-parameter Richard's growth model (Seber and Wild 1989), with only three free parameters as the parameters that measure the asymptotic value was fixed to 1; the point of inflection was parameterized by two parameters: one parameter that measured the timing of the point of inflection (x axis) and a second parameter that measured at what proportion of sexually mature worms the curve started to become concave (y axis). The third parameter measured the slope of the increase in the number of sexually mature worms. Statistical inferences about the time of sexual maturity were made using likelihood ratio tests. Richard's growth models were also used to describe the cumulative cocoon and hatchling productions; the number of cocoons or hatchlings was assumed to be Poisson distributed, where lambda (t) is the mean number of cocoons or hatchlings at time t. Lambda (t) was modeled as a four-parameter Richard's growth model as described above, except all parameters were free. Statistical inferences about differences in cumulative cocoon or hatchling production were made using likelihood ratio tests. Fisher's exact test was used to test if there was a significant difference in number of cocoon-producing couples between the three treatments. Potential differences in cocoon hatching success were tested using Excel Chitest. The correlation between number of cocoons produced per couple and the weight of each couple before (week 39) and after (week 66) cocoon production was assessed using the Pearson's product-moment correlation coefficient in R commander (Fox 2005). Difference in worm riboflavin content was analyzed using the non-parametric Wilcoxon signed rank test in R commander (Fox 2005). RESULTS AND DISCUSSION Establishment and validation of differential earthworm lines The presence of Verminephrobacter, Ca. Nephrothrix and (sporadically, see below) the Agromyces-like symbionts was confirmed by fluorescence in situ hybridization (FISH, Fig. S1, Supporting Information) in the E. andrei control culture, which was started from locally collected compost worms. Kanamycin treatment of E. andrei cocoons resulted in a Verminephrobacter-free worm line (as proven by FISH, Fig. S1, Supporting Information), while treatment with both kanamycin and erythromycin resulted in a fully aposymbiotic worm line (as proven by FISH). The Agromyces-like symbiont was detected by FISH and specific PCR in both the control and the Verminephrobacter-free worm line but not in all individuals; PCR confirmed it in two out of five control individuals, and three out of five Verminephrobacter-free individuals. FISH results indicated a very low abundance (i.e. a few cells per nephridium) and it was not detected in all individuals. These observations are consistent with Davidson et al. (2010), who also reported sporadic presence and low numbers of these symbionts in the nephridia. At such low and sporadic abundance, we consider it unlikely that the Agromyces-like symbiont contributes to the fitness effects measured in our experiment, which we therefore ascribe to the Verminephrobacter and Ca. Nephrothrix symbionts. Nephridial symbionts affect sexual maturation but not growth of their E. andrei host Growth and reproduction of the three worm lines (control, Verminephrobacter-free, aposymbiotic) was compared over a complete E. andrei life cycle under nutrient limitation; this condition was chosen because the fitness benefit of Verminephrobacter in the common garden worm A. tuberculata was only apparent under nutrient limitation (Lund et al.2010b). Consistent with this earlier study (Lund et al.2010b), growth of the worms during the first 39 weeks (measured as weight gain) was not affected by the removal of Verminephrobacter or by the removal of all three symbionts (Fig. 1). Resource limitation was apparent by the drastically prolonged generation times: while well-fed E. andrei can commonly grow to adulthood and develop clitella in only 5 weeks (Reinecke and Viljoen 1991; Domínguez, Briones and Mato 1997), visible tubercula pubertatis and thus sexual maturity in our experiment was first detected after 27–39 weeks (Fig. 2). The time of sexual maturation was slightly delayed but not significantly different from the control in the Verminephrobacter-free E. andrei line (Fig. 2); this is different from A. tuberculata, where removal of Verminephrobacter significantly delayed sexual maturation (Lund et al.2010b). Apparently, Ca. Nephrothrix (and possibly the Agromyces-like symbiont) can partially compensate for the loss of Verminephrobacter, as only the complete removal of the symbionts had a significant effect on sexual maturation (Richard's growth model, P < 0.0001; Fig. 2). Thus, both Verminephrobacter and Ca. Nephrothrix (and the Agromyces-like symbiont) affect the timing of sexual maturation in a cumulative or synergistic manner. Figure 1. View largeDownload slide Growth of E. andrei during 39 weeks and the final weight of the worms in week 66 (after cocoon production). Bars represent averages with standard deviations (n = 20) for control, Verminephrobacter-free and aposymbiotic worms, according to legend. Figure 1. View largeDownload slide Growth of E. andrei during 39 weeks and the final weight of the worms in week 66 (after cocoon production). Bars represent averages with standard deviations (n = 20) for control, Verminephrobacter-free and aposymbiotic worms, according to legend. Figure 2. View largeDownload slide Number of sexually mature worms over time in each group. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 2. View largeDownload slide Number of sexually mature worms over time in each group. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Verminephrobacter but not Ca. Nephrothrix benefit cocoon and offspring production The cumulative cocoon and hatchling production is shown in Fig. 3; the control worms produced significantly more cocoons and hatchlings than both the Verminephrobacter-free and the aposymbiotic worms (Richard's growth model, P < 0.0001: Fig. 3). The hatching success of the cocoons ranged from 58% to 81% but was not significantly different between the treatments (Chi test, P = 0.155; Fig. 4). These combined results indicate that Ca. Nephrothrix had no effect (neither beneficial nor detrimental) on cocoon production or hatching success, while Verminephrobacter is beneficial for E. andrei reproduction by increasing cocoon and thereby hatchling numbers. Verminephrobacter also benefitted reproduction in A. tuberculata; however, here cocoon hatching success (57% in symbiotic vs 25% in aposymbiotic worms) and not cocoon production was affected (Lund et al.2010b). It is unknown why Verminephrobacter affects cocoon production and hatching success differently in the two earthworm species. One explanation could be that hatching success in E. andrei is affected by the number of times the worm has mated (Porto, Velando and Domínguez 2012); if a worm only mated once, the hatching success was 59 ± 11% (same range as this study, Fig. 4), whereas worms mated two to six times had a cocoon hatching success of 80 ± 5%. We did not control the number of matings per worm, which may influence, and thereby obscure, a potential effect of Verminephrobacter on cocoon hatching success in E. andrei. It has, to our knowledge, never been investigated if the number of mating events also influences hatching success in A. tuberculata. Figure 3. View largeDownload slide Cumulative production of cocoons (A) and hatchlings (B) in control, Verminephrobacter-free and aposymbiotic worms. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 3. View largeDownload slide Cumulative production of cocoons (A) and hatchlings (B) in control, Verminephrobacter-free and aposymbiotic worms. The curves represent non-linear regressions using Richard's growth model to describe the data. Different superscript letters in the legend indicate significant differences between the fitted growth models (P < 0.0001). Figure 4. View largeDownload slide Cocoon production and hatching success. Gray and white bars show number of hatched and non-hatched cocoons, respectively. Numbers above bars show hatching success in %. There was no significant difference in hatching success between the three treatments (Chi test, P = 0.155). Figure 4. View largeDownload slide Cocoon production and hatching success. Gray and white bars show number of hatched and non-hatched cocoons, respectively. Numbers above bars show hatching success in %. There was no significant difference in hatching success between the three treatments (Chi test, P = 0.155). Even though the growth of E. andrei was not directly affected by the symbionts (Fig. 1), worms in the control group were able to reproduce at a smaller size than both Verminephrobacter-free and aposymbiotic worms (Fig. 5). If the Verminephrobacter-free and aposymbiotic worms were below a weight threshold of about 0.2 g in week 39, they never produced cocoons, whereas the control worms below this threshold produced one to four cocoons. Control worms consistently produced more cocoons even though they had a similar weight as the Verminephrobacter-free and aposymbiotic worms. Overall, the weight of the worms at the time of mating (week 39) positively correlated with the number of cocoons the couples would produce during the following 27 weeks (Pearson's r = 0.6923, 0.8480 and 0.7417 for the control, Verminephrobacter-free and aposymbiotic groups, respectively, P < 0.005; Fig. 5). The ability of control worms to produce more offspring at a smaller size suggests that Verminephrobacter (but not Ca. Nephrothrix) provide a nutritional supplement (e.g. vitamins) to their host, which allows them to invest in offspring early on. Figure 5. View largeDownload slide Number of cocoons produced per couple as a function of the weight of each worm in week 39. Worms were mated in pairs of two and each couple produced x cocoons. If several couples from the same treatment produced the same number of cocoons, the couples can be differentiated according to the color; two individuals from the same couple have the same color. There is a significant correlation between the weight of the worms and how many cocoons they will produce during the following 27 weeks (control; blue line, Pearson r = 0.6923. Verminephrobacter-free; red line, Pearson r = 0.8480. Aposymbiotic; green line, Pearson r = 0.7417. P < 0.005). If the Verminephrobacter-free and aposymbiotic groups are below a weight threshold of about 0.2 g in week 39 (vertical gray line), they never produce cocoons, whereas the control worms below this threshold produce one to four cocoons. Figure 5. View largeDownload slide Number of cocoons produced per couple as a function of the weight of each worm in week 39. Worms were mated in pairs of two and each couple produced x cocoons. If several couples from the same treatment produced the same number of cocoons, the couples can be differentiated according to the color; two individuals from the same couple have the same color. There is a significant correlation between the weight of the worms and how many cocoons they will produce during the following 27 weeks (control; blue line, Pearson r = 0.6923. Verminephrobacter-free; red line, Pearson r = 0.8480. Aposymbiotic; green line, Pearson r = 0.7417. P < 0.005). If the Verminephrobacter-free and aposymbiotic groups are below a weight threshold of about 0.2 g in week 39 (vertical gray line), they never produce cocoons, whereas the control worms below this threshold produce one to four cocoons. Verminephrobacter symbionts do not alter the riboflavin content of their hosts Further indications for a symbiotic nutrient supplementation that in part compensates for a poor diet are the observation that (i) nutrient limitation was required to detect a fitness effect of Verminephrobacter in A. tuberculata (Lund et al.2010b); (ii) aposymbiotic stock cultures of both E. andrei and A. tuberculata are reproducing normally under nutrient-rich conditions in the lab; and (iii) the fitness effect of the symbionts was similar in both worm species. One possibility for such a supplementation is riboflavin, as recently hypothesized for E. andrei (Sulik et al.2012). We found that the content of riboflavin in cocoons collected from worm stock cultures (not starved) did not differ between the control, Verminephrobacter-free or aposymbiotic worms (Wilcoxon signed-rank test, P = 0.25) (Fig. S2, Supporting Information). Similarly, there were no differences in riboflavin content between worm lines at the end of the fitness experiment (Wilcoxon signed-rank test, P > 0.6, Fig. S3, Supporting Information). Consequently, as Verminephrobacter does not affect earthworm riboflavin content, it is unlikely that it provides the host with this vitamin. CONCLUSION AND FUTURE PERSPECTIVES The collective findings from this study and Lund et al. (2010b) show that in both the compost worm E. andrei and the garden worm A. tuberculata the Verminephrobacter symbionts enhance host reproduction, and that both worm species reach sexual maturity earlier when all of their native symbionts are present. Intriguingly, these beneficial effects appear conserved between host species despite the different ecology and feeding behavior of these earthworms. As these effects are seen under nutrient-limiting conditions, nutritional supplementation appears an obvious explanation. Yet neither the original amino acid recycling proposal (Pandazis 1931) nor the more recent riboflavin hypothesis (Sulik et al.2012) could be confirmed in our comparative fitness studies (Lund et al.2010b; this study). Since the nutritional status of the worm is still likely to be a determinant for the onset of sexual maturity and reproductive success, the question thus remains: What compound(s) do the symbionts provide for their host? Or do they rather detoxify waste products that otherwise weaken the host? Alternatively, the symbionts may also directly manipulate the host reproductive hormone system to ensure early and prolific reproduction, thereby benefitting their own transfer to the next generation. Interestingly, the receptor for the hormone anneotcin, which induces egg-laying behavior in earthworms (Oumi et al.1996), is produced in the nephridia of the clitellum region (Kawada et al.2004), thereby co-localizing with the nephridial symbionts. Manipulation of host reproductive hormones, however, does not explain why the fitness effect is seen under nutrient limitation. Thus, the molecular mechanism behind the fitness benefit of nephridial earthworm symbionts is still to be discovered. SUPPLEMENTARY DATA Supplementary data are available at FEMSEC online. Acknowledgements We thank Annemarie Højmark, Susanne Nielsen and Karina Bomholt Oest for excellent technical assistance and Martin Holmstrup for good advice and fruitful discussions on earthworm ecology. FUNDING This work was funded by the EU Marie Curie Initial Training Network (ITN) Symbiomics (Contract Number 264774), the Aarhus Institute of Advanced Studies (AIAS), and the Danish Council for Independent Research | Natural Sciences (FNU). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. Conflict of interest. None declared. REFERENCES Andrés-Lacueva C, Mattivi F, Tonon D. Determination of riboflavin, flavin mononucleotide and flavin–adenine dinucleotide in wine and other beverages by high performance liquid chromatography with fluorescence detection. J Chromatogr A  1998; 823: 355– 63. Google Scholar CrossRef Search ADS PubMed  Ankrah NYD, Luan J, Douglas AE. Cooperative metabolism in a three-partner insect-bacterial symbiosis revealed by metabolic modeling. J Bacteriol  2017, doi:10.1128/JB.00872-16. Daims H, Brühl A, Amann R et al.   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  1999; 22: 434– 44. Google Scholar CrossRef Search ADS PubMed  Davidson SK, Powell RJ, James S. A global survey of the bacteria within earthworm nephridia. Mol Phylogenet Evol  2013; 67: 188– 200. Google Scholar CrossRef Search ADS PubMed  Davidson SK, Powell RJ, Stahl DA. Transmission of a bacterial consortium in Eisenia fetida egg capsules. Environ Microbiol  2010; 12: 2277– 88. Google Scholar PubMed  Davidson SK, Stahl D. Transmission of nephridial bacteria of the earthworm Eisenia fetida. Appl Environ Microb  2006; 72: 769– 75. Google Scholar CrossRef Search ADS   Domínguez J, Briones M, Mato S. Effect on the diet on growth and reproduction of Eisenia andrei (Oligochaeta, Lumbricidae). Pedobiologia  1997; 41: 566– 76. Douglas AE. The molecular basis of bacterial-insect symbiosis. J Mol Biol  2014; 426: 3830– 7. Google Scholar CrossRef Search ADS PubMed  Douglas AE. How multi-partner endosymbiosesfunction. Nat Rev Microbiol  2016; 14: 731– 43. Google Scholar CrossRef Search ADS PubMed  Dulla G, Go R, Stahl D et al.   Verminephrobacter eiseniae type IV pili and flagella are required to colonize earthworm nephridia. ISME J  2012; 6: 1166– 75. Google Scholar CrossRef Search ADS PubMed  Fox J. The R commander: A basic-statistics graphical user interface to R. J Stat Softw  2005; 14: 1– 42. Kawada T, Kanda A, Minakata H et al.   Identification of a novel receptor for an invertebrate oxytocin/vasopressin superfamily peptide: molecular and functional evolution of the oxytocin/vasopressin superfamily. Biochem J  2004; 382: 231– 7. Google Scholar CrossRef Search ADS PubMed  Lund MB, Davidson SK, Holmstrup M et al.   Diversity and host specificity of the Verminephrobacter-earthworm symbiosis. Environ Microbiol  2010a; 12: 2142– 51. Lund MB, Holmstrup M, Lomstein B et al.   Beneficial effect of Verminephrobacter nephridial symbionts on the fitness of the earthworm Aporrectodea tuberculata. Appl Environ Microb  2010b; 76: 4738– 43. Google Scholar CrossRef Search ADS   Lund MB, Kjeldsen KU, Schramm A. The earthworm-Verminephrobacter symbiosis: an emerging experimental system to study extracellular symbiosis. Front Microbiol  2014; 5: 128. Google Scholar CrossRef Search ADS PubMed  Manz W, Amann R, Ludwig W et al.   Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria—problems and solutions. Syst Appl Microbiol  1992; 15: 593– 600 Google Scholar CrossRef Search ADS   Martín R, Miquel S, Ulmer J et al.   Gut ecosystem: how microbes help us. Benef Microbes  2014; 5: 219– 33. Google Scholar CrossRef Search ADS PubMed  Møller P, Lund MB, Schramm A. Evolution of the tripartite symbiosis between earthworms, Verminephrobacter and Flexibacter-like bacteria. Front Microbiol  2015; 6: 529 Google Scholar PubMed  Oumi T, Ukena K, Matsushima O et al.   Annetocin, an annelid oxytocin-related peptide, induces egg-laying behavior in the earthworm, Eisenia foetida. J Exp Zool  1996; 276: 151– 6. Google Scholar CrossRef Search ADS PubMed  Pandazis G. Zur Frage der Bakteriensymbiose bei Oligochäten. Zentlbl Bakteriol  1931; 120: 440– 53. Porto PG, Velando A, Domínguez J. Multiple mating increases cocoon hatching success in the earthworm Eisenia andrei (Oligochaeta: Lumbricidae). Biol J Linnean Soc  2012; 107: 175– 81. Google Scholar CrossRef Search ADS   Reinecke AJ, Viljoen SA. A comparison of the biology of Eisenia fetida and Eisenia andrei (Oligochaeta). Biol Fertil Soils  1991; 11: 295– 300. Google Scholar CrossRef Search ADS   Schramm A, Davidson SK, Dodsworth J et al.   Acidovorax-like symbionts in the nephridia of earthworms. Environ Microbiol  2003; 5: 804– 9. Google Scholar CrossRef Search ADS PubMed  Schweitzer B, Huber I, Amann R et al.   α-and β-Proteobacteria control the consumption and release of amino acids on lake snow aggregates. Appl Environ Microb  2001; 67: 632– 45. Google Scholar CrossRef Search ADS   Seber GAF, Wild CJ. Nonlinear Regression . New York: John Wiley & Sons, Inc., 1989. Google Scholar CrossRef Search ADS   Sims RW, Gerard BM. Earthworms . Leiden: J. Brill/Dr. W. Backhuys, 1985. Sulik P, Klimek M, Talik P et al.   Searching for external sources of the riboflavin stored in earthworm eleocytes. Invertebr Surv J  2012; 9: 169– 77. © FEMS 2017. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

FEMS Microbiology EcologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off