Present address: R. Kortet, Neurobiology, Physiology and Behavior, University of California, Davis, One Shields Avenue, Davis, CA 95616, USA.
Jari Ahtiainen, Department of Biological and Environmental Science, University of Jyväskylä, P.O. Box 35, FIN-40351, Jyväskylä, Finland. Tel.: +358-14-2601211; fax: +358-14-2602321; e-mail: email@example.com
The field of ecological immunology is ultimately seeking to address the question ‘Why is there variation in immune function?’ Here, we provide experimental evidence that costs of ubiquitous sexual signals are a significant source of variation in immune function. In the mating season, males of the wolf spider Hygrolycosa rubrofasciata drum against dry leaves while wandering around the habitat searching for receptive females. According to a previous study, the male metabolic rate during the drumming increases 22-fold compared to the resting metabolic rate. In the present study, we examined whether investment in costly courtship drumming decreases male immune function in a wild population of H. rubrofasciata. We induced males to increase their drumming rate by introducing females in proximity. As estimates of male immune function, we used lytic activity and encapsulation rate. Lytic activity estimates the concentration of antimicrobial peptides in haemolymph, which have been shown to play an important role in defence against bacteria, viruses and fungi. Encapsulation is an important defence mechanism against nematodes and insect parasitoids, but it also plays a role in defence against viruses. Our results show that males with nonarbitrarily increased investment in drumming rate had considerably lower lytic activities than control males. Also, there was a tendency for males with nonarbitrarily increased investment in drumming rate to have lower encapsulation rates than control males. This study provides experimental evidence for the first time, to our knowledge, that there are direct immunological costs of sexual signalling in natural populations. Therefore, immunological costs of sexual signals may provide significant phenotypic variation to parasite-mediated sexual selection.
Sexual signals are ubiquitous, widely studied and an important currency in sexual selection (Andersson, 1994). To provide honesty for signal receivers, signalling theory assumes costs to signal emitters (Bradbury & Vehrencamp, 1998). The immuno-competence handicap hypothesis provides a mechanism for signalling costs: males with high signalling levels must impose higher immunological costs than males with low signalling levels (Wedekind & Folstad, 1994). As finite resources must be allocated between immune function and sexual signalling (see Roff, 1992), this forms a trade-off that may maintain continuing susceptibility to parasites. Parasite susceptibility may also be maintained by the co-evolutionary arms race between signalers and signal exploiters, as sexual signals have been shown to inadvertently attract parasites (Cade, 1975). However, we do not know of any studies, which have examined the trade-off between sexual signalling and immune function directly. Previous studies have manipulated overall reproductive effort by exposing males to different numbers of females (McKean & Nunney, 2001; Kilpimaa et al., 2003; Fedorka et al., 2004). However, in those experiments, males with increased reproductive effort have experienced all the stages in the species’ courtship behaviour, making it impossible to estimate the effect of sexual signalling on immunity.
Although the direct evidence for the trade-off between sexual signalling and immune function is still missing, there have been many studies which strongly suggest that this kind of trade-off truly exists. These studies have not manipulated sexual signalling per se, but have focused on the trade-off between reproduction and immune function. Thus far, the best evidence for that trade-off has come from studies of resource allocation. The resource allocation hypothesis states that, if available resources limit both reproduction and immune function, investment in either activity should lead to a reduction in the other activity (Sheldon & Verhulst, 1996). One way to look at this hypothesis has been to compare immune function before and after reproduction (Adamo et al., 2001; Kortet et al., 2003). The second way has been to form experimental groups, which, with all other things being equal, have been treated differently in terms of investment in reproduction, after which researchers measured immune response from those groups (McKean & Nunney, 2001; Rolff & Siva-Jothy, 2002; Kilpimaa et al., 2003; Fedorka et al., 2004). The third way has been to manipulate investment in immune function directly, which, in the case of a trade-off, is expected to change investment in reproduction accordingly. There are at least two possibilities of doing this: administration of immuno-suppressive or immuno-enhancing drugs or the use of lines artificially selected for high and low immune function (Verhulst et al., 1999; Norris & Evans, 2000). The fourth way has yielded only indirect evidence by manipulating the resource availability itself (e.g. condition or essential nutrients) (Folstad et al., 1994; Rantala et al., 2003a). The fifth way has also been indirect, examining dual effects of antigen injections on reproduction and immune function (Saino et al., 1997; Westneat et al., 2003). However, the trade-off between reproduction and immune function could also be the consequence of dual effects of hormones (e.g. testosterone or juvenile hormone) on reproduction and immune function without any variation in resource allocation. Indeed, there is an accumulating amount of experimental evidence for the hormone-mediated immuno-competence handicap hypothesis (Olsson et al., 2000; Rolff & Siva-Jothy, 2002; Rantala et al., 2003b).
Invertebrate immune function is based both on humoral and cellular components, which interact with each other in defending against multicellular parasites (insect parasitoids, nematodes) and pathogens (bacteria, viruses, fungi) (Brey & Hultmark, 1998; Lavine & Strand, 2002). Humoral immune function includes the rapid production of a number of small antimicrobial proteins and peptides as well as the activation of the prophenoloxidase (PPO) cascade leading to melanization (Brey & Hultmark, 1998). The main synthesis of induced antibacterial peptides takes place in the fat body – an adipose tissue with somewhat analogous metabolic function to the mammalian liver (Gillespie et al., 1997). Some hemocytes are also involved in the production of antibacterial peptides. In encapsulation response, circulating hemocytes recognize an object as foreign and form an insulating capsule around it (Pathak, 1993; Gillespie et al., 1997). A forming capsule is covered by the black pigment, melanin, which is the result of the PPO cascade. The encapsulated organism faces several killing factors including asphyxiation, production of toxic quinones or hydroquinones via the PPO cascade, reactive oxygen and nitrogen intermediates, and antibacterial peptides (Salt, 1970; Gillespie et al., 1997; Nappi et al., 2000).
In this study, we experimentally demonstrate that investment in costly drumming decreases male immune function in a wild population of the wolf spider H. rubrofasciata. In H. rubrofasciata, costs of male sexual signalling can be tested by manipulating not the trait directly, but the expected pay-off (i.e. probability of mating) by introducing females in proximity (Mappes et al., 1996; Kotiaho, 2000). This increases male signalling rate in a nonarbitrary manner. After the experiment of differential sexual investment, we use lytic activity and encapsulation rate in measuring different aspects of male immune function. Lytic activity estimates the concentration of antimicrobial peptides in haemolymph, which have been shown to play an important role in defence against bacteria, viruses and fungi (review in Bulet et al., 2004). Encapsulation is an important defence mechanism against nematodes and parasitoids (e.g. Gillespie et al., 1997), but it also plays a role in defence against viruses (Washburn et al., 1996). Recently, Ahtiainen et al. (2004) have shown that, without female proximity, male drumming rate is positively associated with male immune function at the population level. There are also many other studies which have shown the positive relationship between male quality and immune function to exist (Rantala et al., 2000,2002,2004; Ryder & Siva-Jothy, 2000; Rantala & Kortet, 2003,2004). However, the present study shows that, at the individual level, H. rubrofasciata males face a trade-off between sexual signalling and immune function.
Materials and methods
Hygrolycosa rubrofasciata (Ohlert, 1865) is a ground-dwelling wolf spider (Araneae: Lycosidae), which can be found in patchily located populations widely distributed over northern Europe. Hygrolycosa rubrofasciata inhabits two kinds of habitats: abandoned fields and other meadow habitats and half-open bogs with deciduous trees (Betula spp., Salix spp.). During the short mating season (April–June), males drum while wandering around the habitat searching for receptive females. Hygrolycosa rubrofasciata males produce drumming signals by hitting their abdomen on dry leaves or other suitable substrates. One drumming consists of ca. 30–40 separate pulses, lasts ca. 1 s, and is audible to the human ear up to a distance of several metres (Rivero et al., 2000). Males drum without the presence of females or female silk but will increase their drumming rate considerably in the presence of a female. If the female is willing to copulate with the male, she responds by vibrating her body. After an initial female response, the male and the female, while approaching each other, drum several times before the copulation begins (the so-called duetting) (Kronestedt, 1996). Hygrolycosa rubrofasciata females prefer males with high drumming rates as mating partners (Kotiaho et al., 1996; Parri et al., 1997; review in Ahtiainen et al., 2001). In H. rubrofasciata, males are polygynous, whereas females seem to copulate only once. Males do not provide nuptial gifs, parental care, or any other obvious direct benefits other than sperm for females. Male drumming incurs substantial physiological costs: during the drumming, the male metabolic rate increases 22-fold compared to the resting metabolic rate (Kotiaho et al., 1998). Males are not found later in the season, and presumably they die after a single mating season due to costs related to sexual signalling.
We collected spiders using pitfall traps from a bog in Sipoo, Southern Finland (60°16′N and 25°14′E) in 3rd and 4th May 2002 at the beginning of the mating season. We placed spiders individually in small plastic film jars with some moss (Sphagnum spp.) and kept them at a cool temperature (ca. +10 °C) in darkness to keep their metabolic rate low. In the laboratory, we weighed spiders to the nearest 0.1 mg with an analytical balance (AND HA-202M). After body mass measurements, we kept spiders individually in moistened film jars filled with some moss (Sphagnum spp.) at +5 ± 2 °C in darkness to keep their metabolic rate low. We provided food (Drosophila melanogaster) ad libitum.
Before we carried out the differential investment experiment, we measured the drumming rate and mobility from H. rubrofasciata males. For drumming rate and mobility measurements, we took each male randomly from the sample of all specimens, and placed them individually in plastic arenas (125 mm × 88 mm × 110 mm high). We covered the bottom of plastic arenas with a piece of white paper (8 cm × 4 cm). To enable mobility measurements, we divided each arena with a line into two equal rectangles. We placed two dry even-sized birch leaves in the rectangles as drumming substrates (one leaf per rectangle). We illuminated the laboratory with fluorescent tubes and lamps with 40 W bulbs placed 30 cm above the floors of plastic arenas to give extra heat and light. On the day prior to behavioural observations, we kept males at the laboratory temperature (+31 ± 1 °C) for two hours to trigger their sexual activity. In terms of microhabitat, H. rubrofasciata prefers elevated grass tussocks (ca. 30 cm × 30 cm) that are covered with dry leaves (Kotiaho et al., 2000). In this microhabitat, effective temperatures may well rise above 30 °C during sunny spring days. During the experiments, we measured drumming rate as the number of separate drumming bouts, and mobility as the number of times the male crossed a line between the rectangles. We measured drumming rate and mobility five times for 2 min during the trial day, and we repeated this procedure on three consecutive days. Between and after the trial days, we fed the males with fruit flies (D. melanogaster) and kept them in moistened film jars filled with some moss (Sphagnum spp.). To keep their activity levels low, spiders were stored at +5 ± 2 °C in darkness. The repeatability (R) for drumming rate and mobility across the trial days was moderate (drumming rate: R = 0.326: F147,296 = 2.452, P < 0.001; mobility R = 0.323: F147,296 = 2.431, P < 0.001).
We assigned males to two different experimental treatments, either high drumming rate treatment (female proximity) or low drumming rate treatment (control), by picking successive males from the systematic rank of male sexual activity to different treatments. In the high drumming rate treatment, we placed one virgin female in a small cage inside the plastic arena (125 mm × 88 mm × 110 mm high) after which we introduced one male to the arena (see Mappes et al., 1996; Kotiaho, 2000). We constructed cages (diameter 2 cm) from cotton netting to allow acoustic and olfactory contact between the male and the female while preventing physical contact. In addition, both sexes could produce and detect vibrations through the paper (8 cm × 4 cm) glued to the bottom of each arena. In the high drumming rate treatment, each female was used only once. In the low drumming rate treatment (control), the female cages were empty, but otherwise we treated the animals similarly. We conducted both treatments simultaneously for12 h in total during two consecutive days (6 h per day). To analyse the difference between treatment groups in the level of sexual signalling, we measured drumming rate and mobility five times for 2 min during the first manipulation day. Between and after the manipulation days, we fed the males with fruit flies (D. melanogaster) and kept them in moistened film jars filled with some moss (Sphagnum spp.) at +5 ± 2 °C in darkness. Then, we assigned males randomly to two different immunological treatments, either the lytic activity or encapsulation rate experiment. Immunological assays were performed within a few days after the experiment of differential sexual investment.
Lytic activity experiment
In the lytic activity experiment, the first set of H. rubrofasciata males was individually CO2-anaesthetized and taped laterally onto separate glass slides. One of the authors (M.R.) took all the haemolymph samples to eliminate the between-observer variation in sampling. A 0.5 μL hemolymph sample was pipetted from a sterile puncture made on the abdomen membrane between the epigastric furrow and the spiracle. This technique was destructive: all the males died as a consequence of hemolymph sampling, preventing us from evaluating the repeatability of individual immune response. The prevalence of parasitoids was checked from each individual, ensuring that no spider was parasitized. The lysozyme activity of the hemolymph was assayed turbidometrically. A hemolymph sample was mixed with 20 μL of phosphate buffered saline solution (PBS, 0.067 M phosphate, 0.9% NaCl, pH 6.4) and frozen at −80 °C. After thawing, vortexed samples were pipetted into a well microplate (Labsystems, Finland). PBS was used as a negative control. Then, samples and controls were mixed with 80 μL of the suspension containing lyophilised cells of Micrococcus lysodeicticus bacterium (Sigma Chemical Co., USA) (0.20 mg mL−1 PBS). To minimize the time difference in the initiation of lytic reaction, bacterial solution was quickly pipetted to each sample (and control). Immediately after that, the optical density at 492 nm was measured at 22 °C in minute intervals on a plate reader (Multiskan Plus, Flow Laboratories, Helsinki, Finland). Lytic activity was expressed as the change in optical density of a sample in the 10 min interval. We acknowledge that future studies should prepare standard curves by using a serial dilution of hen egg white lysozyme. This could allow absolute values of lysozyme to be assigned enabling a better comparison between assays within a species and across species.
Encapsulation rate experiment
In the encapsulation rate experiment, the second set of H. rubrofasciata males were individually CO2-anaesthetized and taped laterally onto separate glass slides. A research assistant made all the implantations. A single sterile 1 mm long piece of nylon monofilament (diameter 0.08 mm) was inserted into the spider's hemocoel through a sterile incision made on the abdomen membrane between the epigastric furrow and the spiracle. The spiders’ immune system was allowed to encapsulate a microfilament for exactly 180 min at constant room temperature (+25 ± 1 °C). The implant was then removed and dried. This technique was destructive: all the males died as a consequence of the removal of the implant. The prevalence of parasitoids was checked for each individual, ensuring that no spider was parasitized. The removed monofilament was examined under a light microscope and recorded on digital video from three different angles. The pictures were then analyzed using an image analysis program (Image Pro). The degree of encapsulation was analysed as a grey value of reflecting light from implants. As a measure of encapsulation rate, we used the average grey values of three video pictures. The scale was calibrated to indicate that the darkest grey received the highest encapsulation rate. To measure the repeatability of this method we re-scanned 16 randomly chosen implants and analysed them as above. The repeatability was very high (R = 0.999: F15,16 = 731.5, P < 0.001).
Before the experiment of differential sexual investment, males did not differ significantly between the treatment groups (i.e. males with or without female proximity) in drumming rate (t146 = 0.315, n.s.), mobility (t146 = 0.300, ns), or body mass (t146 = −0.953, n.s.).
The introduction of a female in the vicinity of a male had a strong positive effect on male drumming rate: males exposed to females drummed 13.7 times more than control males (t146 = −10.015, P < 0.001). However, female vicinity did not have any effect on male mobility (t146 = -0.533, ns), or male mortality (logistic regression enter method, for the model: = 1.159, n.s.).
Males with increased investment in drumming rate had a very significantly lower lytic enzyme activities in their hemolymph than control males (t70 = 4.491, P < 0.001) (Fig. 1). Moreover, there was a tendency for males with increased investment in drumming rate to have lower encapsulation responses against a nylon monofilament than control males (t74 = 1.834, P =0.071) (Fig. 2).
We calculated the within-group correlations between drumming rate measured before the experiment of differential sexual investment, drumming rate measured during the experiment, lytic activity and encapsulation rate. However, this did not reveal any significant patterns.
In the present study, H. rubrofasciata males with nonarbitrarily increased investment in drumming rate had considerably lower lytic activities than control males (Fig. 1). This provides experimental evidence for the first time, to our knowledge, that there are direct immunological costs of sexual signalling in natural populations. Many previous studies have shown that there is a trade-off between reproduction and immune function (see Introduction), but these studies have not manipulated sexual signalling per se. By inducing males to strikingly increase their sexual signalling during courtship, H. rubrofasciata females might inevitably weaken male parasite resistance. Therefore, this sexually induced immunosuppression might provide an opportunity for parasites to exploit. Furthermore, the immunological costs of sexual signals may provide significant phenotypic variation to parasite-mediated sexual selection.
In invertebrates, the fat body synthesizes antimicrobial peptides (Gillespie et al., 1997). Therefore, it is possible that increased investment in energetically costly drumming might have reduced the fat body in H. rubrofasciata males, reducing males’ ability to synthesize lytic enzyme. We suggest that the fat body might work as a proximate mechanism mediating resource allocation between sexual signalling and immune function. However, Svensson et al. (1998) have hypothesized that any trade-off between immune defence and energetically costly activities may not necessarily be based on energy or nutrient limitation but instead may occur either through the adaptive avoidance of autoimmunity or via the damaging effects of free oxygen radical production.
Our results also show that there was a tendency for H. rubrofasciata males with nonarbitrarily increased investment in drumming rate to have lower encapsulation rates than control males (Fig. 2). The encapsulation ability of synthetic objects has been shown to be associated with the encapsulation ability of real parasites (Paskewitz & Riehle, 1994; Gorman et al., 1996). In H. rubrofasciata, encapsulation of a monofilament introduced into a spider's abdomen may well mimic the real situation in the field, as H. rubrofasciata is frequently parasitized by at least two parasitoids, a nematode (Mermithidae: Aranimermis spp.) and a fly (Acroceridae: Ogcodes pallipes; data not shown). These parasitoids cause reproductive failure and eventually the death of infected individuals. Interestingly, encapsulation response has been shown to be positively related to fat content (Koskimäki et al., 2004). This suggests that fat content might be a key physiological mechanism-mediating variation in invertebrate immunity. In the previous study (Ahtiainen et al., 2004), the immune system of H. rubrofasciata males was allowed to encapsulate a microfilament for 240 min, which had been preliminary tested to yield considerable between-individual variation in encapsulation rate. In that study, males with higher drumming rate had higher encapsulation rate at the population level. There was no correlation between drumming rate and lytic activity. In the present study, we used an encapsulation time of 180 min. Therefore, it is possible that the present encapsulation time was not long enough to create sufficient between-male variation in encapsulation rate, which was manifested as a nonsignificant result in the encapsulation rate experiment.
Drumming has been shown to play a crucial role in mate choice of H. rubrofasciata: we have not observed matings to occur without male drumming (Alatalo et al., 1998). Moreover, females mate substantially more with males that have high drumming rates than would be expected if females were passively choosing males in direct proportion to male drumming rates (Kotiaho et al., 1996). Therefore, an increase in male drumming rate has considerable reproductive benefits in terms of increased mating success. However, any decrease in immune function linked to increased drumming rate will potentially enhance the transmission success of phonotactic parasitoids, sexually transmitted diseases and dormant pathogens. In H. rubrofasciata, there is considerable among-male variability in drumming rates without female proximity (Kotiaho et al., 1996). In addition, the male increases his drumming rate more than 10-fold in the presence of the female, and even more if the female responds willingly to the male's mating attempt. Therefore, there is a very substantial amount of phenotypic variability in male drumming rates during the mating season. The present study shows that the variability in acoustic signalling rate creates significant variation in male immune function. Therefore, the trade-off between sexual signalling and immune function can generate evolutionarily significant variation in immunity.
We thank T. Ketola, S. Koistinen, J. Koskimäki, M. Känkänen, J. Tuusa, L. Vainio and J. Valkonen for assistance in the laboratory. This study was funded by the Academy of Finland.