|Home | About | Journals | Submit | Contact Us | Français|
Almost all lumbricid earthworms (Oligochaeta: Lumbricidae) harbor species-specific Verminephrobacter (Betaproteobacteria) symbionts in their nephridia (excretory organs). The function of the symbiosis, and whether the symbionts have a beneficial effect on their earthworm host, is unknown; however, the symbionts have been hypothesized to enhance nitrogen retention in earthworms. The effect of Verminephrobacter on the life history traits of the earthworm Aporrectodea tuberculata (Eisen) was investigated by comparing the growth, development, and fecundity of worms with and without symbionts given high (cow dung)- and low (straw)-nutrient diets. There were no differences in worm growth or the number of cocoons produced by symbiotic and aposymbiotic worms. Worms with Verminephrobacter symbionts reached sexual maturity earlier and had higher cocoon hatching success than worms cured of their symbionts when grown on the low-nutrient diet. Thus, Verminephrobacter nephridial symbionts do have a beneficial effect on their earthworm host. Cocoons with and without symbionts did not significantly differ in total organic carbon, total nitrogen, or total hydrolyzable amino acid content, which strongly questions the hypothesized role of the symbionts in nitrogen recycling for the host.
Symbiosis has long been recognized as a source of evolutionary innovation (24), and the acquisition of symbionts can enable animal hosts to exploit previously inaccessible niches (3). The phylum Annelida is no exception to this; chemosynthetic symbionts in marine annelids (e.g., the giant tubeworm Riftia sp. and other gutless marine oligochaetes ) gain energy from the oxidation of reduced sulfur compounds and fix CO2 and supply their animal host with fixed carbon. A more obscure partnership is known from the bone-eating annelid Osedax sp., where endosymbionts help degrade the bones of whale carcasses, the only known habitat of the worms (33). The medicinal leech Hirudo sp., like other blood-feeding animals (3, 7), has symbionts that are thought to produce essential vitamins missing from a blood meal (13). In addition, leeches have a number of symbionts of unknown function in their nephridia (excretory organs) (18). Earthworms (Oligochaeta: Lumbricidae) have also long been known to harbor symbiotic bacteria in their nephridia (19, 36). The function of this symbiosis, however, is still not known, but the stability of the symbiosis over evolutionary time (23) suggests that the symbionts benefit the host.
The earthworm symbionts reside in the nephridia and have therefore been proposed to be involved in internal recycling of nitrogen in the host (29). The earthworm nephridia play an important role in both nitrogenous waste excretion and osmoregulation (20). The nephridia are found in pairs in each segment of the worm and consist of a coiled tube leading from the coelom to the exterior (Fig. (Fig.1).1). The tube forms three loops, and the symbiotic bacteria are situated in the ampulla in the second loop, where they form a dense population lining the lumen wall (19, 36).
The symbionts form the monophyletic genus Verminephrobacter (Betaproteobacteria) (30, 36); they are species specific and present in almost all lumbricid earthworms (23). The Verminephrobacter symbionts are transmitted vertically via the cocoon, where they are deposited along with eggs and sperm (5). During embryogenesis, the symbionts migrate into the developing nephridia, and after the worm hatches, the symbionts can no longer infect (5, 6). By taking advantage of the vertical transmission mode, it has been possible to establish symbiont-free earthworm cultures in the laboratory through controlled antibiotic treatment of newly deposited cocoons (5; this study). Separation of the symbiotic partners allows studies of the effect of the symbionts on their earthworm hosts.
Pandazis (29) hypothesized that the symbionts enhance earthworm nitrogen retention by excreting proteolytic enzymes that will degrade peptides and proteins lost in urine; this would allow the earthworm to reabsorb the resulting amino acids. As a consequence, earthworms cured of their symbionts should have a lower fitness level than control worms when grown under nitrogen-limiting conditions. To test this hypothesis, growth, development, fecundity, and cocoon hatching success were compared for symbiotic and aposymbiotic earthworms of the species Aporrectodea tuberculata (Eisen) under high and low nutrient availability conditions.
The earthworm A. tuberculata was used as a model organism. It is a medium-size, endogeic worm that feeds on soil organic matter and highly fragmented plant litter (2). A. tuberculata has Verminephrobacter as the only nephridial symbiont (23).
Two earthworm cultures, with and without symbionts, were established. Adult individuals of A. tuberculata were collected from garden soil in Aarhus, Denmark, in October 2005. The worms were kept at 15°C in 1-liter containers with soil supplemented with ground cow dung. The bedding soil was exchanged every 3rd to 5th week. Both the soil and the cow dung were dried at 80°C for 24 h prior to use to kill all indigenous fauna. Cocoons were harvested by wet sieving the bedding soil through a 2-mm mesh. Cocoons were used either to raise control worms or to establish an aposymbiotic worm culture.
Worms were cured of their symbionts by soaking cocoons in antibiotics as previously described (5). Newly laid cocoons (0 to 3 weeks old) were washed thoroughly in tap water and placed in petri dishes with filter paper moistened with kanamycin sulfate (130 to 150 μg·ml−1). The filter paper and kanamycin solution were changed daily. After 15 days of antibiotic treatment, the cocoons were transferred to fresh petri dishes with filter paper and tap water. Cocoons for raising control worms were also kept in petri dishes with tap water; the filter paper was changed weekly until the cocoons hatched. Cocoons were kept at 15°C at all times. Hatchlings from both the control and kanamycin sulfate-treated cocoons were raised to maturity and allowed to breed. To avoid any effect of the antibiotic treatment itself, only worms of the second generation, which had never been in contact with antibiotics, were used in the fitness study.
Individual worms from both the symbiotic and aposymbiotic cultures were checked by fluorescence in situ hybridization (FISH) for the respective presence or absence of Verminephrobacter symbionts and other bacteria. A combination of the probes LSB145-CY3, specific for Verminephrobacter and some Acidovorax species (37), and EUB338-MIX-CY5, targeting almost all bacteria (4), was used. FISH was performed as described in reference 23.
Symbiotic and aposymbiotic worms were also checked for the presence of “Candidatus Lumbricincola,” a novel lineage of Mollicutes reported to be associated with lumbricid earthworms (26). DNA was extracted from the coelomic fluid and body wall of both symbiotic and aposymbiotic worms and used for PCR detection of “Candidatus Lumbricincola” with the specific primers Lum803F and Lum1279R according to reference 25. “Candidatus Lumbricincola” was not detected in any of the earthworm cultures.
Growth, time to sexual maturity, and cocoon production were monitored in symbiotic and aposymbiotic worms subjected to two different diets; one group was fed cow dung to represent a high-nutrient diet, and the other group was fed straw to represent a low-nutrient diet. There were 20 worms in each of the four groups. The total nitrogen (TN) and total organic carbon (TOC) contents of cow dung and straw were determined to be 1.6 mmol N·g−1 and 37.0 mmol C·g−1 for cow dung and 0.3 mmol N·g−1 and 39.4 mmol C·g−1 for straw, resulting in C/N ratios of 23.6 and 122.6 for cow dung and straw, respectively.
A standard food mixture was prepared by mixing 1 volume cow dung or straw with 1 volume soil (1:1 dry volume), and water equal to 50% of the dry weight was added to the mixture. Both the cow dung and the straw had been dried at 80°C for 24 h and finely ground. The food mixture was thoroughly mixed with dried (80°C for 24 h) bedding soil, and water equal to 18% of the dry weight was added. The amount of food given during the experiment was adjusted as the worms grew. For cow dung treatment, worms received a food mixture equal to 1% of the dry weight of soil in weeks 0 to 18 and 3% in weeks 21 to 27; in week 30, the worms were paired and received a 6% mixture. For straw treatment, worms received 1% in weeks 0 to 18 and 0.5% in weeks 21 to 36; in week 39, the worms were paired and received 1%. For the first 6 weeks, hatchlings were given 100 g moist soil-food mixture; this amount was gradually increased to 300 g by week 18 and given throughout the rest of the experiment. Worms were kept at 15°C for the entire experiment.
Before the experiment was initiated, cocoons produced by symbiotic and aposymbiotic worms were harvested and incubated in petri dishes with moist filter paper at 20°C for almost 2 months. All of the cocoons hatched over a period of 9 days, and the hatchlings were kept in petri dishes with moist filter paper until the experiment was initiated. To ensure an equal initial weight distribution in the two food treatments, the hatchlings (symbiotic and aposymbiotic) were arranged by weight and divided into blocks of two. The two members of each block were randomly allocated to one of the two treatments.
During the experiment, worms were weighed every 3rd week and bedding soil with food was exchanged. The worms were washed in tap water and blotted with filter paper before weighing. After 30 weeks, worms receiving cow dung had reached sexual maturity and were randomly paired. Worms receiving straw were paired after 39 weeks, when they had all reached sexual maturity. Cocoons produced by the paired worms were harvested by wet sieving the bedding soil through a 2-mm mesh. Cocoons were incubated in petri dishes with moist filter paper at 15°C until hatching. Cocoons that did not hatch within a 4-month period were recorded as not hatched.
At the end of the experiment, it was verified by FISH (as described above) in worms from the cow dung treatment that all symbiotic worms did have Verminephrobacter symbionts and that all aposymbiotic worms did not have any detectable bacteria in their nephridia.
Worms were allowed to produce cocoons for measurement of cocoon nutrient content. To ensure nutrient limitation in the adult worms, symbiotic and aposymbiotic worms (40 of each) were kept on a straw diet for 3 months before cocoons were harvested. The diet consisted of 400 g bedding soil supplemented with a 1.5% straw mixture for four worms in one pot. The cocoons were harvested and the bedding soil exchanged every 2nd week. The worms receiving cow dung (both symbiotic and aposymbiotic, 40 of each) were also kept in pots with 4 worms in each and they were given a large surplus of cow dung every 2nd week.
TN and TOC contents were determined in 10 freeze-dried individual cocoons from each treatment in a CE Instruments NA2000 elemental analyzer.
Content of total hydrolyzable amino acids (THAA) was determined in five or six freeze-dried individual cocoons (~10 mg [dry weight]) from each treatment. The cocoons were hydrolyzed in 200 μl 6 N HCl at 105°C for 24 h under N2, after which the samples were placed in an ice bath to stop the hydrolysis. Subsamples of hydrolysate (50 μl) were transferred to new glass vials, dried under vacuum at 50°C, resuspended in Milli-Q water, and dried again following the procedure described in reference 22. The dried samples were then dissolved in 4 ml Milli-Q water and filtered (0.2-μm filter; Sartorius) into Pico vials. Before analysis, samples were diluted in Milli-Q water and the concentration of THAA in the hydrolysate was analyzed by reverse-phase high-performance liquid chromatography of fluorescent o-phthaldialdehyde-derivatized products according to the method of Lindroth and Mopper (21). The concentrations of the individual amino acids were calculated from individual standard curves produced from a standard amino acid solution (AA-S-18; Sigma-Aldrich) to which were added β-alanine, γ-aminobutyric acid, taurine, and ornithine. l-Amino butyric acid was used as an internal standard in both standards and samples, as l-amino butyric acid was not present in the samples. Blanks were prepared in the same way as samples and showed negligible concentrations of amino acids from handling and reagents.
The time of sexual maturity was fitted to a flexible Richards 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 three-parameter Richards growth model (38), where 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.
Difference in cocoon hatching success was demonstrated using the Excel Chitest. The chemical composition of cocoons was analyzed by two-way analysis of variance.
Over a period of 48 weeks after hatching, earthworms fed cow dung (high nutrient content) grew faster than worms fed straw (low nutrient content). However, there was no difference in growth between symbiotic and aposymbiotic worms within either of the two feeding groups (Fig. (Fig.22).
Worms receiving cow dung reached sexual maturity about 6 weeks earlier than worms receiving straw (Fig. (Fig.3),3), and there was no difference between symbiotic and aposymbiotic worms on the cow dung diet. In contrast, the symbiotic worms on the straw diet reached sexual maturity significantly earlier than aposymbiotic worms (Fig. (Fig.33).
Worms fed cow dung produced twice as many cocoons as worms fed straw, and the hatching success of these cocoons (84.1 and 78.8% for symbiotic and aposymbiotic worms, respectively) was significantly higher than the hatching success of cocoons from the worms undergoing the straw treatment (Fig. (Fig.4).4). There was no difference in the number of cocoons produced or in hatching success between symbiotic and aposymbiotic worms undergoing the cow dung treatment. In the straw treatment, the hatching success of cocoons produced by symbiotic worms (57.4%) was significantly higher than the hatching success of cocoons produced by aposymbiotic worms (25.0%) (Fig. (Fig.4).4). There was no difference between the symbiotic and aposymbiotic worms in the number of cocoons produced.
Newly laid cocoons (<2 weeks old) from symbiotic and aposymbiotic worms given high- and low-nutrient diets were analyzed for TOC, TN, and THAA contents. There were no differences in TOC or TN content in the cocoons from the four treatments (Table (Table1).1). Cocoons from the four treatments also had the same relative amino acid composition (see Fig. S1 in the supplemental material), and the total amino acid content was not significantly different among the four treatments (Table (Table1).1). More than 57% of the cocoon TOC content and more than 68% of the TN content could be explained by the measured amino acids (Table (Table11).
Symbioses between animals and microbes are diverse and widespread (1, 3), and knowing the fitness effect of symbionts is often essential for understanding the symbiotic associations. In some symbiotic systems, the fitness effect is very clear; if symbionts are removed from sap-, wood-, or blood-feeding insects, an immediate and severe host fitness decrease is seen, because the hosts cannot survive without essential nutrients provided by the symbionts (8, 16, 28). The fitness effect in other symbiotic systems can be more difficult to reveal; in isopods (12), the survival of symbiotic and aposymbiotic individuals only differed on a low-nutrient diet. And in an entomopathogenic nematode, the beneficial effect of the symbiont could only be seen when the host was infecting prey; during the free-living part of the nematode life cycle, the symbiont decreased host fitness (10). The earthworm-Verminephrobacter symbiosis is another example of the difficulty in demonstrating the effect of a symbiont; only under nutrient limitation (as proven by the slower growth [Fig. [Fig.2],2], delayed sexual maturity [Fig. [Fig.3],3], and poor reproduction [Fig. [Fig.4]4] of the straw-fed worms) was there a clear beneficial effect of the symbionts, which was most pronounced during the development of offspring. Nevertheless, this effect is apparently sufficient to explain why the symbionts have been maintained in lumbricid earthworms over evolutionary time (23), even though aposymbiotic worms can both grow and reproduce and even though the symbionts have no apparent effect on host fitness under nutrient-rich conditions. Knowing during which part of its life cycle and under which conditions a host is most dependent on its symbionts may be pivotal in elucidating the function of the symbiosis.
The N retention hypothesis (29) predicts symbiotic worms to have higher internal recycling of nitrogenous compounds lost in urine than aposymbiotic worms: the role of the symbionts is to excrete proteolytic enzymes that degrade small proteins lost in urine, thereby enabling the host to reabsorb the resulting amino acids.
The putative higher nitrogen availability in symbiotic worms could explain the earlier sexual maturity detected in symbiotic worms growing on straw (Fig. (Fig.3).3). Also, the difference in cocoon hatching success (Fig. (Fig.4)4) could be explained by the N retention hypothesis: if symbiotic worms recycle nitrogen, they will be able to invest more nutrients in individual cocoons, thereby increasing hatching success. This would also explain the higher hatching success of cocoons produced by worms living on the high-nutrient cow dung diet (Fig. (Fig.4).4). However, the lack of correlation between cocoon nutrient content (Table (Table1)1) and hatching success (Fig. (Fig.4)4) contradicts Pandazis' N retention hypothesis. In addition, pure cultures of “Verminephrobacter tuberculatae” (the native symbiont of A. tuberculata) are unable to hydrolyze proteins extracellularly (M. B. Lund et al., unpublished data) and the genome of the closely related species Verminephrobacter eiseniae, the nephridial symbiont of Eisenia fetida, does not encode any known extracellular proteases (JGI project 4001428; N. Pinel et al., unpublished data). These collective findings all reject the N retention hypothesis.
Surprisingly, there were no differences in TOC, TN, or THAA content of newly produced cocoons from the different treatments, not even between the cow dung and straw treatments (Table (Table1).1). These components could therefore not explain the differences in cocoon hatching success (Fig. (Fig.4).4). The THAA made up 44 to 50% of the cocoon dry weight (Table (Table1),1), which is slightly lower than the total protein content of E. fetida cocoons (58%) (31). We only determined 15 of the 20 protein amino acids, which may explain the lower amino acid content. The amino acid content was measured in whole cocoons, whereby a part of the amino acids originates from the cocoon wall. Cocoon walls are composed of keratin-like proteins (25), and it is unknown how much these proteins contribute to the total THAA content. The measured amino acids accounted for more than 57% of the TOC and more than 68% of the TN in A. tuberculata cocoons (Table (Table1).1). This means that only a small fraction of the carbon and nitrogen present can be bound in sugars and lipids. The sugar and lipid contents of cocoons from A. tuberculata are unknown, but newly produced cocoons of E. fetida contain about 30% hydrolyzable sugars and about 10% lipids (32). The small amounts of sugars and lipids in cocoons make it unlikely that variations in these compounds can explain the differences in cocoon hatching success between symbiotic and aposymbiotic worms and between the cow dung and straw treatments.
Due to their pronounced effect on cocoon hatching success, the Verminephrobacter symbionts are likely to play an important role during embryogenesis and possibly a lesser role in juvenile and adult worms. Three alternative hypotheses of symbiont function during different worm life stages will be discussed below: (i) detoxification of host excretory products in both embryos and juvenile/adult worms, (ii) protection of the host against pathogenic soil bacteria especially during embryogenesis, and (iii) vitamin production for the host, which may be of particular importance during embryogenesis.
In pure culture, V. eiseniae can utilize both ammonia and urea as sole nitrogen sources (30); the symbionts may therefore be involved in the detoxification of nitrogenous waste excreted by worms or by developing embryos. Various invertebrate symbionts utilize host nitrogenous waste; this includes endosymbionts of gutless marine oligochaetes that have been implicated in the detoxification and recycling of host nitrogenous waste (40) and uric acid utilization by the symbionts of brown planthoppers (35), shield bugs (17), and molgulid tunicates (34). The Blochmannia endosymbionts in carpenter ants degrade urea and incorporate the resulting ammonia into amino acids that are utilized by the host (11).
Detoxification of urea and ammonia may be important in juvenile and adult worms during prolonged periods of inactivity (aestivation) and starvation. Little is known about earthworm nitrogen excretion during aestivation; whole-body urea content rises significantly, but the worms do not switch to ureotelism (1a, 15). Starvation, on the other hand, induces a switch from ammonotelism to ureotelism (27, 39). The mechanism behind these changes in nitrogen waste products is unknown, but the earthworm symbionts may consume, and thereby detoxify, both ammonia and urea. An example of nitrogen recycling by symbionts is the uric acid recycling in shield bugs that is crucial for host survival during diapause (17). During the earthworm growth experiment, aestivation happened more frequently in worms receiving the low-nutrient diet than in well-fed worms (worms would empty their intestines and curl up in a ball [data not shown]). If the symbionts are primarily affecting their hosts when the worms are aestivating, the higher aestivation frequency in the starved, straw-fed worms might explain why the symbiont effect on the time of sexual maturity was only observed in this group (Fig. (Fig.3).3). However, the experiment was not designed to investigate the effect of symbionts on earthworm aestivation, and the earlier sexual maturity of juvenile symbiotic worms could also be explained by symbiont amino acid or vitamin production.
During embryogenesis, the symbiont consumption of host excretory products may also be important, explaining the higher hatching success of symbiotic cocoons (Fig. (Fig.4).4). The lack of difference in amino acid content among the four treatments (Table (Table1;1; see Fig. S1 in the supplemental material) implies that the symbionts are not providing the embryos with amino acids; thus, the mechanism may be purely for detoxification. Although congruent with most other data, the detoxification hypothesis fails to explain why cocoons from the cow dung treatment had significantly greater hatching success than cocoons from the straw treatment (Fig. (Fig.44).
Alternatively, the function of the Verminephrobacter symbionts could be to protect the developing embryo from pathogenic soil bacteria, thereby increasing cocoon hatching success. However, known antimicrobial compounds have not been identified in the genome of V. eiseniae (JGI project 4001428; Pinel et al., unpublished), and therefore symbiont-mediated protection commonly found in other symbiotic systems (14) seems unlikely. Although the possibility that genes of unknown function (13.4% of V. eiseniae genes) encode antimicrobials cannot be ruled out, the symbiont-mediated protection hypothesis still fails to explain the significantly higher hatching success of cocoons from the cow dung versus the straw treatment (Fig. (Fig.4)4) and the earlier sexual maturation of symbiotic worms (Fig. (Fig.33).
A third function of the Verminephrobacter symbionts could be to supply their earthworm hosts with essential vitamins and cofactors; vitamin supplementation may be of particular importance during embryogenesis, when the embryo is restricted to feeding on the nutritious mucus in the cocoon. In contrast, adult earthworms continuously ingest soil microbes, which may provide the essential vitamins and cofactors. The only vitamin needed for V. eiseniae growth is biotin (30), and the full genome of V. eiseniae encodes all other vitamin synthesis pathways (JGI project 4001428; Pinel et al., unpublished). The majority of the genes in these pathways have also been found in the partial genome of “V. tuberculatae subsp. tuberculatae,” the native symbiont of A. tuberculata (K. U. Kjeldsen et al., unpublished data). The dependency of insect hosts on vitamin supplements from their symbionts is known from a number of symbioses (3, 7). The higher hatching success of cocoons from the cow dung treatment (Fig. (Fig.4)4) can be explained if the cocoon vitamin content affects embryo viability. Worms from the cow dung treatment may deliver more vitamins to their cocoons, whereas worms from the straw treatment may depend on their symbionts for vitamin production in the cocoons. These small differences were not detectable in the cocoon C and N analyses, as vitamin content was not directly measured. Vitamin synthesis by the symbionts may even be of importance in adult worms, as shown by the earlier sexual maturity of symbiotic than aposymbiotic nutrient-limited worms (Fig. (Fig.33).
In conclusion, the Verminephrobacter symbionts have a beneficial effect on host fitness, as evident from the increased cocoon hatching success and earlier time of sexual maturity of symbiotic worms. However, the positive effect was only seen when the earthworms were nutrient limited. The N retention hypothesis formulated by Pandazis in 1931 (29) is not supported by the presented data or by the findings in studies of pure Verminephrobacter cultures (Lund et al., unpublished; 30). We propose the symbionts to be important for either (i) vitamin biosynthesis, which fits the current data best, or (ii) detoxification of nitrogenous waste, particularly during embryogenesis. These two alternative hypotheses require further investigation by examining cocoon vitamin content and urea and ammonia contents. These results underline the importance of relevant experimental conditions and patience by the experimenter when elucidating the effect of diverse and widespread bacterial symbionts on the fitness of their respective hosts.
We thank Kasper Urup Kjeldsen, Ditte Carlsen, Susanne Juhler, and Daniel Aagren Nielsen for helping taking care of the earthworms. Thanks to Lykke Poulsen and Bodil Andersen for valuable technical assistance.
This study was financially supported by the Danish Research Council for Natural Sciences (grant 21-04-0410 to A.S.).
Published ahead of print on 28 May 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.