Bdelloid rotifers are aquatic microinvertebrates common in water bodies and in unstable “terrestrial” habitats, such as mosses and lichens. The key to the adaptability to live in unstable habitats is their capacity to tolerate habitat desiccation through anhydrobiosis, that is assumed apomorphic to the taxon. The life history traits of some “moss” and “water” species of bdelloid are compared, showing that the water species have shorter life span, higher fecundity and earlier age at first reproduction than the moss species. These traits are discussed in the light of current life history theories. Contrary to the assumptions of the models, anhydrobiosis of bdelloids does not appear to imply energy demand. Past research on bdelloid anhydrobiosis is briefly reviewed, focusing on the factors that affect anhydrobiosis success, like morphological and physiological adjustments, and on the effect of events of anhydrobiosis during life time. Desiccation produces a time shift on the age of the bdelloid, which disregards the time spent as anhydrobiotic, following the so-called “Sleeping Beauty” model. Average fecundity is never found to decrease as a consequence of anhydrobiosis, but is either equal or even higher than that of a hydrated rotifer. Bdelloid populations seem to benefit from anhydrobiosis; fitness of a bdelloid is found to decline, if populations are maintained hydrated for several generations, but not if populations are cyclically desiccated. We hypothesize that anhydrobiosis can be an essential event for long-term survival of bdelloid populations.
ROTIFERS AND ANHYDROBIOSIS
Sediments of lotic and lentic waters, as well as the thin water film around soil particles, mosses and lichens are common habitats of Bdelloid rotifers. These are microscopic aquatic invertebrates that are widely distributed and are well adapted to unstable habitats, mainly because of two biological characteristics (Ricci, 1987). (1) Bdelloids reproduce through obligatory apomictic (ameiotic) parthenogenesis, only. After catastrophic events, parthenogenesis allows them to elude the risks connected to low population density, and increases the potential reproductive rate of the population under suitable conditions. (2) Bdelloids are capable of anhydrobiosis, a form of dormancy triggered by the loss of water due to evaporation. This strategy allows them to escape the hazard of a temporally unfavourable habitat, and to be ready to re-establish when conditions improve.
A bdelloid, on perceiving that water is evaporating, undergoes a series of morphological and physiological adjustments in preparation to anhydrobiosis. It contracts into a compact shape, called “tun,” loses water from tissues and from the major body cavity and reduces its weight to about 30–40% of the hydrated condition (Dickson and Mercer, 1967; Ricci et al., 2003, and, C. R., unpublished data). Differently from most organisms capable of anhydrobiosis (Clegg, 2001), bdelloids do not synthesize trehalose as a protective chemical, while other rotifers belonging to class Monogononta possess trehalose in the dormant stage, an arrested embryo called “resting egg” (Lapinski and Tunnacliffe, 2003; Caprioli et al., 2004). The production of the resting egg requires a complex cascade of reproductive events that are initiated by some remote cue, its dormancy is temporarily irreversible and is broken by specific triggers, not necessarily linked to the cessation of the unfavourable condition (Gilbert, 1974; Pourriot and Snell, 1987; Schroeder, 2005). It does not represent the response to a rapidly changing habitat, as monogonont rotifers commonly occupy habitats that become unsuitable cyclically (“coarse-grained” environments, sensuLevins, 1968; see also Ricci, 2001a). In contrast, bdelloids occupy unstable and unpredictable habitats exposed to frequent desiccation, where conditions quickly switch from suitable to unfavourable. The key to their adaptability is the rapidity with which they can enter and recover from the anhydrobiotic state.
ANHYDROBIOSIS AND BDELLOID SPECIES
Most species of bdelloids are capable of anhydrobiosis (Ricci, 1998), but some, scattered over the four bdelloid families, were found non resistant to desiccation. It seems unlikely that species that are phylogenetically related to each other have acquired the capacity of anhydrobiosis independently of each other, while it is more parsimonious to think that all bdelloids have inherited the desiccation resistance from a common ancestor, and that some of them have secondarily lost it. Therefore the anhydrobiotic capability is assumed to be apomorphic to the taxon Bdelloidea (Ricci, 1998). Nevertheless, the probability to face desiccation differs among species living in different environments, either water bodies or terrestrial mosses, that dry out more or less frequently. Thus the life-history traits of different species is expected to be affected by the desiccation probability and frequency in their habitat, to some extent (see also Jönsson and Järemo, 2003).
If we divide bdelloids on the basis of habitat, substantial differences of life history traits distinguish “water” and “moss” bdelloid species (Table 1). Of course, the values here reported were obtained in the lab and cannot be readily extended to natural populations, but somehow reflect the characteristics of the different species. Although statistical comparisons of the life-table traits of the species would be not appropriate because the data were obtained under non-comparable conditions, the aquatic species had shorter life span and higher fecundity than the moss-dwelling species, and matured at an earlier age. There is no evidence to state that the longer life span of moss species has to be related to slower body growth, or to reduced expenditures in reproduction. This set of traits, low fecundity, delayed maturity, and long life span is instead predicted by Jönsson and Järemo (2003) as the result of trade-offs between anhydrobiotic capacity and reproductive investments, under the hypothesis that anhydrobiosis implies costs. Actually, a delayed maturity reduces the cost of reproduction and increases adult survival (Stearns and Crandall, 1981), and these traits are common to animals living in uncertain environments (Shaffer, 1974). In addition, the same set of traits, low reproduction and long lifetime, characterizes animals adapted to habitats where adversity selection (A-selection) operates (Whittaker, 1975; Greenslade, 1983). Moss bdelloid species do possess these traits, but, while A-selection is expected to occur in habitats that undergo unfavourable conditions predictably (Greenslade, 1983; Parsons, 1994), conditions of terrestrial mosses, soil and lichens appear to switch from suitable to unsuitable unpredictably. These habitats have been defined ‘fine-grained’ for bdelloids because each animal can be exposed to several changes during its life time (Ricci, 2001a). In response to the unpredictability of habitat disturbance, they enter anhydrobiosis quickly, at any age, without necessitating specific ontogenetic stages, and break dormancy as soon as conditions become suitable again.
Dormancy of bdelloids is a fast response to disturbance through a direct on-off switch, without the delayed response or the bet-hedging strategy that characterizes the dormancy of monogonont rotifers. While the production of the resting egg is very costly to monogononts, it is not clear whether anhydrobiosis imposes costs on the bdelloids. Although it is known that they do not synthesize trehalose, it is still unclear what molecule is used. Late Embryogenesis Abundant (LEA) proteins are known to be produced in desiccation resistant plants (Ingram and Bartels, 1996; Clegg, 2001) have been found in nematodes (Browne et al., 2002) and are good candidates for bdelloid rotifers, as well (A. Tunnacliffe, personal communication). Experimental results suggest that bdelloids do not use stored resources to survive desiccation (Santo et al., 2001; Ricci et al., 2004), in spite of some contrasting evidence (Lapinski and Tunnacliffe, 2003).
Is there any cost to a bdelloid entering and emerging from dormancy? Costs of anhydrobiosis for bdelloids can be measured in terms of mortality, affected by the animal's age (Örstan, 1995; Orsenigo et al., 1998; Ricci, 1998), food or starvation (Lapinski and Tunnacliffe, 2003; Ricci et al., 2004), and condition during desiccation, like nature of substrate or duration of drought (Ricci et al., 2003).
If dormancy imposes costs, animals that face successive anhydrobiosis will recover in lower number. Alternatively, if anhydrobiosis favours the fittest animals, all those emerging from the first anhydrobiosis will be able to survive the second anhydrobiosis, as well, and a higher recovery should be expected after the second anhydrobiosis. To test these hypotheses, under lab conditions, cohorts of a bdelloid species (Macrotrachela quadricornifera) were desiccated when 8-d-old for one week, re-hydrated, desiccated again one week later for 7 days, and then re-hydrated to measure their recovery rates. Although the age of the rotifers at first and second anhydrobiosis differs, and rotifer age is known to affect anhydrobiotic capacity, no significant difference was found between the fraction of rotifers emerging from the first (recovery = 99.3%) and the second (recovery = 94.4%) anhydrobiosis, according to Pearson's Chi-squared test with simulated P-value based on 2000 replicates (chi-squared = 4.36, df = NA, P-value = 0.050). On this basis, we can assert that 1) a bdelloid species survives more events of desiccation during its life time, 2) the probability to recovery after each anhydrobiosis does not vary.
ANHYDROBIOSIS AND LIFE HISTORY SCHEDULE
At each desiccation every rotifer enters into dormancy more or less simultaneously and at each re-hydration every rotifer breaks dormancy according to an on-off switch. Although anhydrobiosis may exert a selective force because not all species are equally able to survive and, on the whole, juveniles may be affected more severely than adults, yet mortality is equally distributed all over the population. At re-hydration, bdelloids will resume normal activity in a matter of few hours and the reproductive animals will resume reproduction. The response of bdelloids to habitat desiccation has been defined a reactive deterministic phenomenon (Ricci, 2001a).
When dormant, animals could either keep on or suspend the aging process. The two alternative options were referred to as “The Picture of Dorian Gray” and “Sleeping Beauty” respectively. To test what option bdelloids follow, cohorts of M. quadricornifera, M. vanoyei and Adineta ricciae were desiccated and rehydrated after a week, and their life history traits after recovery were compared to those of hydrated controls (Ricci et al., 1987; Ricci and Abbruzzese, 1989; Ricci and Covino, 2005). No difference of mean life span was found between desiccated and hydrated cohorts, if the days spent dry were not considered in the desiccated animals' age (Table 2). Species were found to ignore the time spent as anhydrobiotic, all resumed reproduction when re-hydrated, and “reset” their age at the start of dormancy, following the “Sleeping Beauty” model. When a different group of animals, a nematode species, that is equally capable of anhydrobiosis was tested to explore its life history after dormancy, it was found to age during the anhydrobiosis period, meeting the predictions of “The Picture of Dorian Gray” model (Ricci and Pagani, 1997). As for age-specific fecundity rates (mean egg number per animal per day) of the bdelloids, M. quadricornifera and M. vanoyei were found to produce the same number of eggs whether desiccated or hydrated, while after anhydrobiosis A. ricciae produced significantly more eggs than its hydrated control, suggesting that this bdelloid species can benefit from desiccation (Table 2). On the basis of these results it can be asserted that the bdelloids investigated so far disregarded the time spent as anhydrobiotic and, for at least one species, the event of anhydrobiosis might trigger some physiological process giving advantages after dormancy. A positive effect on animals which will re-start a population to occupy an empty environment might be reasonable in terms of adaptation, but what ‘mechanisms’ operate is an open question.
ANHYDROBIOSIS AND BDELLOID POPULATIONS
Some authors (Dobers, 1915; Hickernell, 1917) claimed that bdelloid populations can degenerate if never desiccated, and Ricci (1987) showed that fitness-related traits (fecundity and generation time) of a bdelloid population (Philodina roseola) declined over years. However, the population could resume the original trait values after emerging from anhydrobiosis. This observation raised the question whether bdelloids require to be exposed to anhydrobiosis to keep fitness at acceptable levels. We wanted to test the hypothesis that bdelloids need to be “stressed” by assessing the difference of fitness between hydrated and desiccated populations. The experimental models were M. quadricornifera, which on recovery keeps fecundity constant, and A. ricciae, which on recovery increases fecundity. For each species were established two sub-populations, one of which was maintained under continuous hydration for 6 months, and the other one was desiccated monthly during 7 days. To investigate the effect of anhydrobiosis on population fitness, life-table experiments were run to record fecundity at beginning (for reference) and after several desiccations, comparing the ‘hydrated’ and the ‘desiccated’ lines to the initial reference value. In both species recovery percentages were constant along the experimental time, fecundity was significantly lower and generation time was longer in the “hydrated” line, while both traits were close to the original figures in the “desiccated” line (Fig. 1). On the whole, these results confirm the observations of Dobers (1915), Hickernell (1917) and Ricci (1987), and point out that populations of three bdelloid species (P. roseola, M. quadricornifera, A. ricciae) ‘need’ to enter anhydrobiosis to avoid decline of fitness. We do not know the cause of the decline, as mutation accumulation is non realistic. Actually there is no evidence of exceptionally high mutation rate in these animals, that are eutelic (nuclei number fixed at birth) in all tissues, including ovaries, produce eggs through mitotic divisions and never mate, because they are parthenogenetic. The expected mutational load due to parthenogenesis cannot be so fast to become effective in 4–5 generations, and more important, cannot be reversible.
We showed the remarkable degradation of bdelloid populations if kept constantly hydrated, but do not have explanations to offer for the favourable effect of anhydrobiosis on these populations. Only hypotheses can be advanced at this stage. We might imagine that the continuous parthenogenetic reproduction, that is obligatory to the bdelloids, causes modifications to their DNA, and that emergence from anhydrobiosis promotes repair mechanisms, that are capable of restoring the integrity and the function of the molecule. Alternatively, we can assume that the bdelloids host viruses or parasites, whose load increases with generations if conditions are suitable, and the load reduces the fitness of the population. If the virus or the parasite are less tolerant to desiccation than the bdelloids themselves, then the desiccation could represent a way to decrease the parasite/virus load and keep animal fitness constant. Something happens during anhydrobiosis that is important for the rotifers and has a beneficial effect, but these are mere speculations for the moment, and deserve investigations.
Bdelloids are naturally capable of anhydrobiosis at any age, and we speculate that their capacity was present prior to their radiation, being apomorphic to the taxon. The habitats where they commonly live operate a continuous selection on them, nevertheless not all species are equally good anhydrobionts, and the higher or lower capacity well reflects the frequency of desiccation they are exposed to. Bdelloid species from habitats that commonly desiccate, like terrestrial mosses, are characterized by longer lifespans, lower fecundity and delayed maturation, meeting the predictions generated by many theoretical models. But, in contrast with assumptions of models, anhydrobiosis of bdelloids does not appear to demand energy. The costs imposed on the animals are paid for in terms of mortality, that is affected by several factors, like animal's age or condition, or by the duration, and not by the number of desiccations. The consequences of desiccation on the life of a bdelloid are insignificant for its life history, because the age of the emerging bdelloid corresponds to its age at the onset of anhydrobiosis, and the duration of drought is removed entirely from its life span (Sleeping Beauty model). Reproduction is restored the day after recovery, and lasts as long as in the hydrated rotifer. Average fecundity is never found to decrease in rotifers recovered from anhydrobiosis. Fecundity after anhydrobiosis is either equal (M. quadricornifera, M. vanoyei) or higher (A. ricciae) than that of the hydrated animal. The “factor(s)” linked to desiccation that is capable of enhancing reproduction is to be assessed. Possibly, the same “factor” operates also on populations whose fitness declines, if they are maintained hydrated for several generations. The decline can be broken if the population is desiccated or the “factor” can never act in populations that are cyclically desiccated. Since the “factor” can be removed, it seems likely that its nature is epigenetic, while mutations are to be ruled out.
On the basis of the whole body of data, bdelloids, although aquatic animals, appear very efficient in tolerating desiccation, well adapted to respond to unpredictable and fast changes of water availability, and somehow dependent on anhydrobiosis, which can even represent an essential event in their life. Thus, life in unpredictable habitats should not be seen as the result of an adaptation, but a requirement for bdelloid long term survival.
From the Symposium Drying Without Dying: The Comparative Mechanisms and Evolution of Desiccation Tolerance in Animals, Microbes, and Plants presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 4-8 January 2005, at San Diego, California.
Many thanks are due to Peter Alpert, James Clegg, Brent Mishler and Mel Olivier for organizing the symposium and inviting us to contribute. Chiara Boschetti, Diego Fontaneto and Giulio Melone helped with technical and experimental support. Two anonymous referees improved clarity with their comments and recommendations. Financial support came from an ASI (Italian Space Agency) grant to C.R.