Spatiotemporal and gender‐specific parasitism in two species of gobiid fish

Abstract Parasitism is considered a major selective force in natural host populations. Infections can decrease host condition and vigour, and potentially influence, for example, host population dynamics and behavior such as mate choice. We studied parasite infections of two common marine fish species, the sand goby (Pomatoschistus minutus) and the common goby (Pomatoschistus microps), in the brackish water Northern Baltic Sea. We were particularly interested in the occurrence of parasite taxa located in central sensory organs, such as eyes, potentially affecting fish behavior and mate choice. We found that both fish species harbored parasite communities dominated by taxa transmitted to fish through aquatic invertebrates. Infections also showed significant spatiotemporal variation. Trematodes in the eyes were very few in some locations, but infection levels were higher among females than males, suggesting differences in exposure or resistance between the sexes. To test between these hypotheses, we experimentally exposed male and female sand gobies to infection with the eye fluke Diplostomum pseudospathaceum. These trials showed that the fish became readily infected and females had higher parasite numbers, supporting higher susceptibility of females. Eye fluke infections also caused high cataract intensities among the fish in the wild. Our results demonstrate the potential of these parasites to influence host condition and visual abilities, which may have significant implications for survival and mate choice in goby populations.

Consequently, the impact of parasites can also vary among host populations and, if persistent, such differences can potentially create different selection pressures for host individuals living in these populations (Karvonen & Seehausen, 2012).
Overall, current evidence strongly suggests that male sexual ornaments could signal resistance to parasitism. However, reduction in host vision could impair the ability of individuals to perceive sexual signals. For example, it has been shown in cichlid fishes that visually perceived sexual signals advertised through male coloration can be blurred because of increased water turbidity, resulting in hybridization of species following the relaxation of color-based sexual selection (Seehausen, van Alphen, & Witte, 1997). Similarly, parasites found in the key sensory organs, such as the eyes, could impair host vision and the ability to perceive visual cues from potential mates (Karvonen & Seehausen, 2012). In species where sexual selection is based on males displaying secondary sexual characteristics, the ability of females to judge male quality when infected with such parasites could be compromised. However, while mating decisions are known to be influenced by general condition (Cotton, Small, & Pomiankowski, 2006) and also parasite infections (Lopez, 1999;Mazzi, 2004;Pfennig & Tinsley, 2002;Poulin & Vickery, 1996) of the choosier sex, the potential of parasites directly interfering with sexual selection operating through visual signals perceived by females has remained virtually unexplored. Here, we explore parasitism and particularly the infections in the eyes of gobiid fishes where females actively choose males based on secondary sexual characteristics.
Gobies (Gobiidae) are abundant fish species living in marine and brackish water habitats around the world. Five species of gobies inhabit the Baltic Sea, two of which, the sand goby (Pomatoschistus minutus, Pallas 1770) and the common goby (Pomatoschistus microps, Krøyer 1838), are the most common. During the reproductive period in early summer, males build nests where they attract females to spawn using secondary sexual traits. In the sand goby, for instance, these include a bright blue spot on the first dorsal fin and the size of the fin itself, as well as specific courtship behavior (Forsgren, 1992;Lindström, St. Mary, & Pampoulie, 2006). Also, the size of the nest is one of the key determinants of reproductive success of a male because the most fecund females cannot lay all their eggs in small nests (Lindström, 1992). All these characteristics of male quality are perceived visually by females.
Previous studies have shown that gobies also harbor a diverse parasite fauna. For example, Baltic gobies can host four to 22 species of parasites depending on the host species and sampling time (Zander, 2003). Further, Zander (2005) reported that the parasite communities were often most diverse in autumn with very few species present in the spring. Studies have also reported higher parasite infection among wild-caught female sand gobies compared to males (Van Damme & Ollevier, 1994), suggesting higher exposure and/or susceptibility of females. Finally, behavioral trials have shown that infection of male sand gobies with macroparasites in the body cavity and on the fins and skin did not affect male dominance or female mate choice (Barber, 2002). However, Barber (2002) also found that infection intensity of Gyrodactylus monogeneans had a negative effect on the development of the dorsal fin size, a potential secondary sexual trait of the sand goby.
We explored parasite infections of the sand goby and the common goby in the Baltic Sea by conducting a replicated sampling campaign of male and female fish to capture spatiotemporal variation and possible gender differences in parasitism. We were particularly interested in variation in infection of parasites inhabiting sensory organs of fish that could show the potential of parasite-induced changes in conditions of sexual selection and mate choice among the sampling locations. Similarly, infections could influence overwinter survival of the fish and result in lower infection levels in the early summer. Furthermore, to explain patterns of infection of the parasites in the wild, we exposed male and female sand gobies to controlled experimental infection from trematode eye flukes in the laboratory. Differences in the abundance of these parasites between male and female fish under similar level of exposure would be consistent with the idea of differences in susceptibility between the sexes.

| Sampling of gobies
Sand gobies and common gobies were sampled from three locations in the proximity of the Tvärminne Zoological station, southern Finland. The first location was next to the station (referred here to as    fish eye lenses were studied before dissection for coverage of cataracts caused by Diplostomum spp. eye flukes using slit-lamp (Kowa SL-15) microscopy (Karvonen et al., 2004a). Cataracts were scored as 10%, 20%,…,100% coverage of the lens area, which correlates with the deleterious effects of the parasites on fish Seppälä, Karvonen, & Valtonen, 2005) and thus provides an indirect measure of parasite-induced effects on the host.
Differences in total parasite abundance between the locations and sampling times were analyzed using GLMs with negative binomial distribution and log link function. Abundances of the parasite taxa Trichodina sp. and Gyrodactylus sp. (see Results) were excluded from the analyses as they were studied only from one pectoral fin.

| Experimental exposure
Experimental exposure of sand gobies ( that there is no detectable population genetic structure in these parasites across Finland (Louhi, Karvonen, Rellstab, & Jokela, 2010), which is why parasite origin was unlikely to affect the results. The snails were allowed to produce cercariae for 2 hr in 2 dl of water (20°C). Suspensions of the snails were then combined, and the cercarial density was estimated by taking ten 1 ml samples. Water in the containers was regularly mixed during the first hour of exposure to ensure equal exposure of all individuals to the parasite. The fish were maintained in these conditions for 18 hr, which is sufficient time for the parasites to reach the eye lenses in a small fish (Louhi, Sundberg, Jokela, & Karvonen, 2015). There was no mortality of fish during or after the exposure. All fish were then euthanized and studied for the number of parasites in the eye lenses. Older infections originating from the wild and those originating from the experimental exposure could be distinguished based on the size of the parasite metacercariae. All experimental procedures were in accordance with the ethical standards of the Finnish Regional State Administrative Agency and conducted under license (License code: ESAVI/4706/04.10.07/2015). Data were analyzed using ANCOVA with fish sex as a fixed factor and container as a random factor to account for dependency among fish exposed in the same container.
Fish length was used as a covariate. All analyses were conducted using SPSS 24 statistical package.

| RE SULTS
In total, 12 parasite taxa were detected among the 200 sand gobies and 168 common gobies examined, including five trematodes, two nematodes, two acanthocephalans, one cestode, one monogenean, and one protozoan (   Figure 3b). Similarly to the field data (see above), there was a significant difference in parasite abundance between the sexes so that females were more heavily infected (F 1,3.0 = 10.56, p = .047 (sex); F 2,23 = 0.58, p = .568 (sex × container); Figure 3b). This suggests higher susceptibility of females to infection. New or recent infections were not detected among the 10 unexposed control fish indicating that there was no natural exposure from the water during the experiment and that the fish had not been recently infected in the wild. Four of the 30 exposed fish and one of the 10 control fish harbored fully developed Diplostomum metacercariae, likely as a result of infection in the wild in the previous summer.

| D ISCUSS I ON
Spatial and temporal variation in parasitism can have significant implications for host populations by underlying variability in influence of parasites on the condition and reproduction of individuals (Lefevre et al., 2009;Tompkins, Dunn, Smith, & Telfer, 2011;Wood et al., 2007). This is particularly clear with infections that can cause detrimental effects on the hosts, such as those located in the central sensory organs. Moreover, hosts species in which such effects can come about already at low parasite numbers may suffer the most notable consequences. We examined parasite infections of two species of gobiid fishes in the Northern Baltic Sea. We scored the entire macroparasite community in these fishes and specifically targeted eye flukes of the genus Diplostomum that are known to cause loss of vision in several fish species, with the effects likely being most prominent in smaller fish species with the smallest eye lenses. We found significant spatiotemporal variation in infections as well as differences between the fish species so that the common gobies were more heavily infected in most cases. While there was no effect of host sex on the infections overall, female sand gobies were more heavily infected with eye flukes at the sampling location with the highest infection. An experimental exposure of the fish indicated that the difference between the sexes was at least partly explained by the higher susceptibility of females to infection.
Spatiotemporal variation in parasitism is a common feature of most host-parasite interactions, including parasitic infections of fish in freshwater (Marcogliese, Gendron, Plante, Fournier, & Cyr, 2006;de Roij & MacColl, 2012) and in sea (Grutter, 1998;Sikkel, Nemeth, McCammon, & Williams, 2009). We found that parasite infections of the two goby species followed similar patterns. First, infection abundances were clearly different between the locations, particularly in the common goby, and mainly driven by the trematode Cryptocotyle sp. This could reflect, for example, spatial variation in infection prevalence in the first intermediate snail hosts of the parasite, which is commonly observed also in other trematode systems (Faltýnková et al., 2008;Jokela & Lively, 1995). Second, with few exceptions, directly transmitted parasites (Trichodina sp. and Gyrodactylus sp.) were clearly more prevalent in the early summer compared to autumn, reflecting the temperature-driven replication of these parasites (Bagge & Valtonen, 1999;Halmetoja, Valtonen, & Taskinen, 1992;Koskivaara, Valtonen, & Prost, 1991;Rintamäki-Kinnunen & Valtonen, 1997). Third, trophically transmitted parasites (cestodes, nematodes, and acanthocephalans) showed sporadic occurrence at low numbers, suggesting that the gobies unlikely are the primary fish hosts for these parasites in this system. Fourth, many of the larval trematodes, particularly in sand gobies from the Vargskär location with the highest infection, tended to be more abundant in autumn compared to early summer (Table 1). This well reflects the typical accumulation of trematodes in their intermediate hosts during summer months (Faltýnková, Karvonen, & Valtonen, 2011;Karvonen, Hudson, Seppälä, & Valtonen, 2004;Karvonen et al., 2004a,b). The lower abundance in the early summer is also consistent with the idea that the most heavily infected individuals may be lost from the population during winter, while our data were too few to test this properly. For example, Diplostomum infections and cataracts comparable to this study in the eye lenses are known to cause serious fitness consequences in fish (Crowden & Broom, 1980;Seppälä et al., 2005), supporting a possibility of parasite-driven population effects (Marcogliese, Compagna, Bergeron, & McLaughlin, 2001). Overall, broad lines of the spatial and temporal variation of parasitism in this system are in accordance with earlier findings of parasite infections in gobies in the Baltic Sea (Zander, 2003(Zander, , 2005Zander & Kesting, 1998;Zander, Strohbach, & Groenewold, 1993).
We also found a distinct difference between male and female sand gobies in the abundance of eye fluke infection so that females were more heavily infected both in the field and in the experimental exposure. This is in contrast with the general pattern of higher infestation in males across a range of host-parasite systems (Klein, 2004;Poulin, 1996), although few studies have reported higher parasite infections also in females, for example, in guppies (Richards, van Oosterhout, & Cable, 2010;Stephenson, van Oosterhout, Mohammed, & Cable, 2015;Tadiri, Scott, & Fussmann, 2016), gobies (Van Damme & Ollevier, 1994) and coral reef fish (Sikkel, Fuller, & Hunte, 2000). One reason for the sex difference in this system could be that males and females in the field are differently exposed to the parasite cercariae. Our sampling time in early summer coincided with the breeding season of gobies (Hesthagen, 1977;Lindström, 1998;Nyman, 1953) when males are defending nests and eggs, and remain mostly stationary. On the other hand, females are actively swimming around in search of spawning and feeding opportunities, which could increase their exposure to the parasite cercariae (see also Sikkel et al. (2000)). However, our exposure experiment showed that the sex difference in infections is not merely due to behavioral differences, but that females are also more susceptible to infection than males. This is interesting as it contradicts with the general trend of males being more susceptible to parasite infections because of sex hormones that can suppress immune function (Klein, 2004).
The reason why such a sex-specific infection pattern was observed only in Vargskär is currently unclear, but may be related to differences between the habitat types. For example, the sampling sites Station and Långholmen are more sheltered compared to the more exposed Vargskär, although this would suggest lower rather than TA B L E 2 Prevalence (% fish infected) and mean abundance (   higher infection in the latter. On the other hand, the population size of seagulls, terns, and merganserids, the definitive hosts for many trematodes including Diplostomum, may be larger at Vargskär than the other sites (Lindström & Ranta, 1992). This could enhance parasite life cycles locally (Marcogliese et al., 2001). However, details of the differences in infection processes need further work.
Our data also show that just one to two worms infecting an eye lens of small fish, such as gobies, can severely harm the lens (i.e., cataracts covering the majority of lens were formed at very low infection). This is because the size of the parasite metacercariae (and the damage they inflict per lens volume) likely remain relatively constant regardless of the lens size, but the size of the eye lens increases with fish size. Thus, in larger fish species, cataract coverage typically increases linearly with the parasite abundance and tens of parasites per lens may be required for high cataract intensities Karvonen et al., 2004a), whereas even a low-level infection is likely to severely impair the vision of a small fish (see also Owen, Barber, and Hart (1993)). During mating, female sand gobies visit several males (Forsgren, 1997;Lindström & Lehtonen, 2013) and base their mate choice on a range of visual cues (Forsgren, 1992;Lindström et al., 2006). As a consequence, attractive males reach high mating success compared to less attractive ones (Lindström & Seppä, 1996) and this nonrandom distribution of mating success results in sexual selection (Andersson, 1994;Emlen & Oring, 1977). This process, however, can be potentially affected by impaired visual ability of females. For example, decreased water clarity interferes with visual abilities of females and this has been invoked as an explanation for weakened sexual selection in turbid waters (Järvenpää & Lindström, 2004;Seehausen et al., 1997). Similarly, if the visual ability of females was hampered by Diplostomum, this could affect the way females can judge males and express their mating preferences, as they would not be able to detect male mating signals. Consequently, mating systems in areas of high infection risk for females could become more random and result in weakened sexual selection. Such a process could easily create spatial variation in the intensity of sexual selection and may contribute to preserving genetic variation in male secondary sexual traits. However, female preferences are also based on the quality of male parental care (Lindström et al., 2006;Pampoulie, Lindström, & St. Mary, 2004). Thus, it is possible that females in areas of high infection, on average, select males exhibiting lower quality care, which then results in lower offspring production. While our data are suggestive of the potential for such parasite-induced changes in mate choice, these questions need to be tackled experimentally.

ACK N OWLED G M ENTS
We thank Tvärminne Zoological Station for logistic support. David Marcogliese gave valuable comments on an earlier draft.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
AK and KL conceived, designed, and performed the experiments. AK analyzed the data. AK and KL wrote the manuscript.